Search

Keywords Assignment; Stable matching; College admission; Distributional constraints; Integer programming Abstract Admission to universities is organised in a centralised scheme in Hungary. In this paper we investigate two major specialities of this application: ties and common quotas. A tie occur when some students have the same score at a programme. If not enough seats are available for the last tied group of applicants at a programme then there are three reasonable policies used in practice: 1) all must be rejected, as in Hungary 2) all can be accepted, as in Chile 3) a lottery decides which students are accepted from this group, as in Ireland. Even though student-optimal stable matchings can be computed efficiently for each of the above three cases, we developed (mixed) integer programming (IP) formulations for solving these problems, and compared the solutions obtained by the three policies for a real instance of the Hungarian application from 2008. In the case of Hungary common quotas arise from the faculty quotas imposed on their programmes and from the national quotas set for state-financed students in each subject. The overlapping structure of common quotas makes the computational problem of finding a stable solution NP-hard, even for strict rankings. In the case of ties and common quotas we propose two reasonable stable solution concepts for the Hungarian and Chilean policies. We developed (mixed) IP formulations for solving these stable matching problems and tested their performance on the large scale real instance from 2008 and also for one from 2009 under two different assumptions. We demonstrate that the most general case is also solvable in practice by IP technique. Author Affiliation: (a) Department of Operations Research and Actuarial Sciences, Corvinus University of Budapest, H-1093, Fovám tér 13-15., Budapest, Hungary (b) Institute of Economics, Research Centre for Economic and Regional Studies, Hungarian Academy of Sciences, H-1097, Tóth Kálmán u. 4., Budapest, Hungary (c) Budapest University of Technology and Economics, H-1111, Muegyetem rakpart 3., Budapest, Hungary * Corresponding author. Article History: Received 28 January 2020; Accepted 19 August 2021 (footnote)[white star] Earlier results of this paper have been presented in two conference papers [5,6]. (footnote)1 Supported by the Hungarian Academy of Sciences under its Momentum Programme (Engineering Economics in Matching Markets, no. LP2021-2), and by the Hungarian Scientific Research Fund -- OTKA (no. K129086). Byline: Kolos Csaba Ágoston [kolos.agoston@uni-corvinus.hu] (a), Péter Biró [peter.biro@krtk.mta.hu] (1,*,a,b), Endre Kováts [endre.kovats.92@gmail.com] (c), Zsuzsanna Jankó [zsuzsanna.janko@uni-corvinus.hu] (a,b)

Keywords Multiple objective programming; Nonsmooth optimization; Semi-quasidifferentiable function; Nonconvex programming Highlights * Introducing semi-quasidifferentiability. * Investigating semi-quasidifferentiable functions and establishing a mean value theorem. * Discussing SQDs of locally Lipschitz functions. * Characterizing normal and adjacent cones in terms of SQDs. * Characterizing weak efficiency in terms of SQDs. Abstract In this paper, we extend the concept of quasidifferential to a new notion called semi-quasidifferential. This generalization is motivated by the convexificator notion. Some important properties of semi-quasidifferentials are established. The relationship between semi-quasidifferentials and the Clarke subdifferential is studied, and a mean value theorem in terms of semi-quasidifferentials is proved. It is shown that this notion is helpful to investigate nonsmooth optimization problems even when the objective and/or constraint functions are discontinuous. Considering a multiobjective optimization problem, a characterization of some cones related to the feasible set is provided. They are used for deriving necessary and sufficient optimality conditions. We close the paper by obtaining optimality conditions in multiobjective optimization in terms of semi-quasidifferentials. Some outcomes of the current work generalize the related results existing in the literature. Author Affiliation: (a) Applied Mathematics Group, Urmia University of Technology, Urmia, Iran (b) School of Mathematics, Institute for Research in Fundamental Sciences (IPM), P.O. Box: 19395-5746, Tehran, Iran (c) School of Mathematics, Statistics and Computer Science, College of Science, University of Tehran, Tehran, Iran * Corresponding author at: Applied Mathematics Group, Urmia University of Technology, Urmia, Iran. Article History: Received 27 November 2020; Accepted 31 October 2021 Byline: Alireza Kabgani [a.kabgani@ipm.ir] (*,a,b), Majid Soleimani-damaneh [soleimani@khayam.ut.ac.ir] (c)

Keywords Scheduling; Parallel Batch Processing; Survey; Semiconductor Manufacturing; Future Research Directions Highlights * We focus on deterministic scheduling of jobs on batch processing machines. * A taxonomy of parallel batching problems is provided. * We survey the literature with parallel batching. * Recent trends are discussed. * Future research directions are identified. Abstract Multiple jobs are processed simultaneously on a given batch processing machine in parallel batching. The resulting batch is called a p-batch. Batching can lead to reduced production costs, but depending how the jobs are grouped into a batch can lead to better or worse delivery times of products. Scheduling jobs on batch processing machines requires grouping decisions in addition to the conventional assignment and sequencing decisions. Parallel batching is important in such diverse areas such as semiconductor manufacturing, aircraft manufacturing, shoe manufacturing, and healthcare. This paper surveys the literature on parallel batching and will focus primarily on deterministic scheduling. We provide a taxonomy of parallel batching problems, distinguishing the compatible case where all jobs can be used to form a batch from the incompatible families setting where only jobs from the same family can be used to form a batch. Makespan, flow time-, and due date-related measures are considered. We discuss scheduling approaches for single machines, parallel machines, and other environments such as flow shops and job shops. In addition to the discussion of archived and current papers, we discuss also recent trends in scheduling jobs on machines with parallel batch processing. Finally, we provide a discussion of future research directions for p-batch scheduling. Author Affiliation: (a) Department of Supply Chain Management, Arizona State University, 85287-5906 Tempe, AZ, USA (b) Department of Mathematics and Computer Science, University of Hagen, 58097 Hagen, Germany * Corresponding author at: University of Hagen, Faculty of Mathematics and Computer Science, Universitätsstraße 1, 58097 Hagen, Germany. Article History: Received 22 July 2019; Accepted 6 June 2021 Byline: John W. Fowler (a), Lars Mönch [lars.moench@fernuni-hagen.de] (b,*)

Keywords Supply chain management, Used apparel collection, Socially responsible operations; Promotion effort Highlights * Comprehensively explore the commercial used apparel collection (UAC) programs. * Analytically derive the optimal promotion effort and study 'profit' coordination. * UAC is better for all members if there're more environmental conscious consumers. * Collecting own branded products will reduce market demand. * Propose a novel and robust effort cost sharing contract to improve UAC. Abstract Motivated by the commonly observed commercial used apparel collection (UAC) programs in fashion retail supply chain operations, this paper analytically explores the associated operational challenges that firms face. First, in the basic model, we consider the case when a fashion retail brand promotes its UAC program and collects the used apparel from consumers. Depending on the conditions of the collected apparel products, the fashion retail brand will classify and either donate them for charity or send them to remanufacturing. For either case, the fashion retail brand gains a benefit. We analytically derive in closed-form the optimal promotion effort and study the mechanism for 'profit' coordination. Our results indicate that many traditional supply chain contracts fail to achieve 'profit' coordination. Thus, the effort cost sharing (ECS) contract is proposed and proven to be effective for 'profit' coordination. Finally, to check the robustness of findings from the basic model, we examine various UAC practice related extended models, such as i) consumer coupon offering, ii) 'no remanufacturing' model, iii) consumer heterogeneity in environmental consciousness, iv) own-brand collection vs any-brand collection. Our findings show that the qualitative insights continue to hold in the extended scenarios. Some novel scenario-dependent findings are also obtained. Insights derived in this paper not only contribute to the literature but also provide scientifically sound managerial guidance and insights to practitioners on how UAC programs can be best operated. Author Affiliation: (a) School of Economics and Management, Southeast University, Nanjing 210093, China (b) Department and Graduate Institute of Business Administration, College of Management, National Taiwan University, Roosevelt Road, Taipei 10617, Taiwan (c) School of Management, Shanghai University, Baoshan, Shanghai, China * Corresponding author. Article History: Received 2 August 2020; Accepted 18 May 2021 (footnote) 'The authors sincerely thank the editor and reviewers for their very helpful and important comments, especially during the very difficult time with COVID19 pandemic. They feel very grateful and hope all editors and reviewers stay safe from the virus.' Tsan-Ming Choi's research is partially supported by The Hong Kong Polytechnic University (Grant Number: G-UA1Q). All authors have good contribution to this paper and the authorship listing follows an alphabetical order of their surnames. Byline: Ya-Jun Cai [julianne.cai@connect.polyu.hk] (a), Tsan-Ming Choi [tmjchoi@gmail.com] (b,*), Ting Zhang [tcheung@shu.edu.cn] (c)

Keywords Data envelopment analysis; Markov manpower planning; Markov processes; Efficiency; Goal programming Highlights * A data envelopment analysis enabled markovian model for decision making. * Radial and additive Markov-DEA models. * Tractable markov manpower planning using data envelopment analysis. * Converting stochastic to deterministic approaches through DEA and goal programming. * A decomposable measure of efficacy of alternative paths to a target outcomes vector. Abstract This paper introduces a modelling framework which combines Data Envelopment Analysis and Markov Chains into an integrated decision aid. Markov Chains are typically used in contexts where a system (e.g. staff profile in a large organisation) is at the start of the planning horizon in a given state, and the aim is to transform the system to a new state by the end of the horizon. The planning horizon can involve several steps and the system transits to a new state after each step. The transition probabilities from one step to the next are influenced by both organisational and external (non-organisational) factors. We develop our generic methodology using as a vehicle the homogeneous Markov manpower planning system. The paper recognizes a gap in existing Markovian manpower planning methods to handle stochasticity and optimization in a more tractable manner and puts forward an approach to harness the power of DEA to fill this gap. In this context, the Decision Maker (DM) can specify potential anticipated future outcomes (e.g. personnel flows) and then use DEA to identify additional feasible courses of action through convexity. These feasible strategies can be evaluated according to the DM's judgement over potential future states of nature and then employed to guide the organisation in making interventions that would affect transition probabilities to improve the probability of attaining the ultimate state desired for the system. The paper includes a numerical illustration of the suggested approach, including data from a manpower planning model previously addressed using classical Markov modelling. Author Affiliation: (a) Quantitative Methods & Decisions Analytics Lab, Department of Business Administration, University of Macedonia, 156 Egnatia Street, 54636 Thessaloniki, Greece (b) Department of Information and Operations Management, Aston Business School, Aston University, Birmingham B4 7ET, United Kingdom (c) Aristotle University of Thessaloniki, Department of Mathematics, Thessaloniki, Greece * Corresponding author. Article History: Received 23 December 2019; Accepted 23 June 2021 Byline: Andreas C. Georgiou [acg@uom.edu.gr] (a,*), Emmanuel Thanassoulis [e.thanassoulis@aston.ac.uk] (b), Alexandra Papadopoulou [apapado@math.auth.gr] (c)

Keywords Global optimization; Quasi-convex feasibility problem; Subgradient method; Iteration complexity; Convergence rate Highlights * A unified framework of subgradient methods is proposed to solve the QFP. * Quantitative convergence theory is established for different stepsizes and controls. * Iteration complexity results validate the benefit of a stochastic control. * Convergence rates are established under the Hölder-type error bound property. * Numerical results show the benefit of QFP and the efficiency of subgradient methods. Abstract The feasibility problem is at the core of the modeling of many problems in various areas, and the quasi-convex function usually provides a precise representation of reality in many fields such as economics, finance and management science. In this paper, we consider the quasi-convex feasibility problem (QFP), that is to find a common point of a family of sublevel sets of quasi-convex functions, and propose a unified framework of subgradient methods for solving the QFP. This paper is contributed to establish the quantitative convergence theory, including the iteration complexity and the convergence rates, of subgradient methods with the constant/dynamic stepsize rules and several general control schemes, including the [alpha]-most violated constraints control, the s-intermittent control and the stochastic control. An interesting finding is disclosed by iteration complexity results that the stochastic control enjoys both advantages of low computational cost requirement and low iteration complexity. More importantly, we introduce a notion of Hölder-type error bound property for the QFP, and use it to establish the linear (or sublinear) convergence rates for subgradient methods to a feasible solution of the QFP. Preliminary numerical results to the multiple Cobb-Douglas productions efficiency problem indicate the powerful modeling capability of the QFP and show the high efficiency and stability of subgradient methods for solving the QFP. Author Affiliation: (a) Shenzhen Key Laboratory of Advanced Machine Learning and Applications, College of Mathematics and Statistics, Shenzhen University, Shenzhen 518060, PR China (b) Department of Mathematics, Statistics and Insurance, The Hang Seng University of Hong Kong, Shatin, Hong Kong (c) Department of Logistics and Maritime Studies, Faculty of Business, The Hong Kong Polytechnic University, Hung Hom, Hong Kong * Corresponding author. Article History: Received 22 April 2020; Accepted 21 September 2021 Byline: Yaohua Hu [mayhhu@szu.edu.cn] (a), Gongnong Li [gnli@szu.edu.cn] (a), Carisa Kwok Wai Yu [carisayu@hsu.edu.hk] (b,*), Tsz Leung Yip [t.l.yip@polyu.edu.hk] (c)

Keywords Routing; Networks; Logistics; Conic programming and interior point methods Highlights * New models for the design of routes without underlying graph structure. * Exact second order cone formulations for crossing postman problem with neighbors. * Matheuristic algorithm for large size instances. * Exhaustive computational experiments on planar graph instances. Abstract In this paper we deal with an extension of the crossing postman problem to design routes that have to visit different shapes of dimensional elements rather than edges. This problem models the design of routes of drones or other vehicles that must visit a number of geographical elements to deliver some good or service and then move directly to the next using straight line displacements. We present two families of mathematical programming formulations. The first one is time-dependent and captures a number of characteristics of real applications at the price of using three indexes variables. The second family of formulations is not time-dependent, instead it uses connectivity properties to ensure the proper definition of routes. We compare them on a testbed of instances with different shapes of elements: second order cone (SOC) representable and polyhedral neighborhoods and polygonal chains. The computational results reported in this paper show that our models are useful and our formulations can solve to optimality medium size instances of sizes similar to other combinatorial problems including neighborhoods that have already been studied in the literature. To address larger instances we also present a heuristic algorithm that runs in two phases: clustering and Variable Neighborhood Search. This algorithm performs very well since it provides promising feasible solutions and, in addition, it can be used to initialize the solvers with feasible solutions. Author Affiliation: Department of Statistics and Operations Research, University of Seville, Seville 41012, Spain * Corresponding author. Article History: Received 18 January 2021; Accepted 29 June 2021 (footnote)1 Both the authors contributed equally to this work. Byline: Justo Puerto [puerto@us.es] (1), Carlos Valverde [cvalverde@us.es] (*,1)

Keywords Finance; Infrastructure investment; Cost-benefit analysis; Real options Highlights * The paper studies the value of abandonment options in projects with construction uncertainty. * The random evolution of construction progress and free cash flows are co-modeled. * The option to abandon construction can generate substantial value and can speed up investment. * A new value-for-money measure is introduced to address specific shortcomings of the benefit-to-cost ratio. Abstract A model of investment in projects with uncertain construction time is presented in which both the cash-flows upon completion and construction progress are modeled as diffusions. The decision maker has three sequential options: (i) to start construction, (ii) to abandon the project during construction, and (ii) to start operations or abandon after completion. Existence of a solution and some of its qualitative properties are proved. A Markov-chain approximation is used to numerically approximate the solution is introduced. The model is applied to a real-world railway project for which it is found that it does not represent value for money (VfM) and is unlikely to do so in the next decade. A comparative statics analysis reveals that the presence of an abandonment option can increase project value substantially and can speed up investment. Finally, it is argued that the commonly-used benefit-to-cost ratio (BCR) is not an appropriate measure to summarize a project's VfM. Instead, the value ratio (VR) is suggested as an alternative. Author Affiliation: Management School & Department of Mathematics, University of York, University Road, York YO10 5DD, United Kingdom Article History: Received 17 April 2020; Accepted 1 July 2021 (footnote)[white star] I am grateful for the research assistance that was provided by Arwyn Goos during an earlier phase of this project. Helpful comments were received from Maria Lavrutich, seminar participants at The York Management School, Aston Business School (Birmingham, UK) and NTNU (Trondheim, Norway), as well as participants of EURO2019 (Dublin, Ireland) and the 23rd Annual Real Options Conference 2019 (London, UK). Support from the Research Council of Norway through project no. 268093 is gratefully acknowledged. I thank three anonymous referees and the editor for their insights and comments, which have greatly improved the paper. Byline: Jacco J.J. Thijssen [jacco.thijssen@york.ac.uk]

Keywords Risk management; Credit scoring; Machine learning; Interpretability; Econometrics, Highlights * We propose a credit scoring model which combines interpretability and forecasting performances. * Our approach is based on an adaptive lasso logistic regression model with predictors extracted from decision trees. * Feature engineering restricts intrinsic complexity of credit score models. * It leads to significant reduction in misclassification costs compared to the benchmark logistic regression. * It is competitive with respect to random forest algorithm. Abstract In the context of credit scoring, ensemble methods based on decision trees, such as the random forest method, provide better classification performance than standard logistic regression models. However, logistic regression remains the benchmark in the credit risk industry mainly because the lack of interpretability of ensemble methods is incompatible with the requirements of financial regulators. In this paper, we propose a high-performance and interpretable credit scoring method called penalised logistic tree regression (PLTR), which uses information from decision trees to improve the performance of logistic regression. Formally, rules extracted from various short-depth decision trees built with original predictive variables are used as predictors in a penalised logistic regression model. PLTR allows us to capture non-linear effects that can arise in credit scoring data while preserving the intrinsic interpretability of the logistic regression model. Monte Carlo simulations and empirical applications using four real credit default datasets show that PLTR predicts credit risk significantly more accurately than logistic regression and compares competitively to the random forest method. Author Affiliation: (a) EconomiX-CNRS, University of Paris Nanterre, 200 Avenue de la République, Nanterre 92000, France (b) Univ. Orléans, CNRS, LEO (FRE 2014), Rue de Blois, Orléans 45067, France * Corresponding author. Article History: Received 13 February 2020; Accepted 23 June 2021 (footnote)[white star] This paper has previously circulated under the title 'Machine Learning or Econometrics for Credit Scoring: Let's Get the Best of Both Worlds'. We are grateful to Emanuele Borgonovo (the editor), three anonymous referees, Jérémy Leymarie, Thomas Raffinot, Benjamin Peeters, Alexandre Girard, and Yannick Lucotte. We also thank the seminar participants at University of Orléans as well as the participants of the 16th Conference 'Développements Récents de l'Econométrie Appliquée à la Finance' (Université Paris Nanterre), 7th PhD Student Conference in International Macroeconomics and Financial Econometrics, 35th Annual Conference of the French Finance Association and International Association for Applied Econometrics for their comments. Finally, we thank the ANR programs MultiRisk (ANR-16-CE26-0015-01), CaliBank (ANR-19-CE26-0002-02), and the Chair ACPR/Risk Foundation: Regulation and Systemic Risk for supporting our research. Byline: Elena Dumitrescu [elena.dumitrescu@parisnanterre.fr] (a), Sullivan Hué [sullivan.hue@univ-orleans.fr] (*,b), Christophe Hurlin [christophe.hurlin@univ-orleans.fr] (b), Sessi Tokpavi [sessi.tokpavi@univ-orleans.fr] (b)

Keywords Robustness and sensitivity analysis; Stability; DEA; Interval scale data; Efficiency measure Highlights * Our paper is about DEA with interval scale data. * More specifically, our focus is on robustness of efficiency scores. * We illustrate our analysis with two numeric examples. Abstract Our paper focuses on a robustness analysis of efficiency scores in the context of Data Envelopment Analysis (DEA) assuming interval scale data, as defined in A. Dehnokhalaji, P. J. Korhonen, M. Köksalan, N. Nasrabadi and J. Wallenius, 'Efficiency Analysis to incorporate interval scale data', European Journal of Operational Research 207 (2), 2010, pp. 1116--1121. We first show that the definition of the efficiency score used in our paper is a well-defined measure according to Aparicio and Pastor (J. Aparicio and J. T. Pastor, 'A well-defined efficiency measure for dealing with closest targets in DEA', Applied Mathematics and Computation 219 (17), 2013, pp. 9142--9154.). Next, we characterize how robust the efficiency scores are with respect to improvements and deteriorations of inputs and outputs. We illustrate our analysis with two examples: a simple numerical example and a more complex example using real-world data. Author Affiliation: (a) Department of Mathematics, Faculty of Mathematical Science and Statistics, University of Birjand, Birjand, Iran (b) Operations and Information Management Department, Aston Business School, Aston University, Birmingham, UK (c) Department of Information and Service Management, Aalto University School of Business, Helsinki, Finland (d) Portsmouth Business School, Centre of Operational Research and Logistics (CORL), University of Portsmouth, United Kingdom * Corresponding author. Article History: Received 9 April 2021; Accepted 22 June 2021 Byline: Nasim Nasrabadi (a), Akram Dehnokhalaji (b), Pekka Korhonen (c), Banu Lokman (d), Jyrki Wallenius [Jyrki.Wallenius@aalto.fi] (*,c)

Keywords OR in banking; Machine learning; Algorithmic fairness; Credit scoring Highlights * A systematic overview of fairness implementation techniques is provided. * The adequacy of established fairness criteria for credit scoring is examined. * Empirical experiments on seven data sets analyze agreement of fairness criteria. * Promising fairness processors are identified based on the empirical results. * Analysis of the Pareto frontiers examines the profit-fairness trade-off. Abstract The rise of algorithmic decision-making has spawned much research on fair machine learning (ML). Financial institutions use ML for building risk scorecards that support a range of credit-related decisions. Yet, the literature on fair ML in credit scoring is scarce. The paper makes three contributions. First, we revisit statistical fairness criteria and examine their adequacy for credit scoring. Second, we catalog algorithmic options for incorporating fairness goals in the ML model development pipeline. Last, we empirically compare different fairness processors in a profit-oriented credit scoring context using real-world data. The empirical results substantiate the evaluation of fairness measures, identify suitable options to implement fair credit scoring, and clarify the profit-fairness trade-off in lending decisions. We find that multiple fairness criteria can be approximately satisfied at once and recommend separation as a proper criterion for measuring the fairness of a scorecard. We also find fair in-processors to deliver a good balance between profit and fairness and show that algorithmic discrimination can be reduced to a reasonable level at a relatively low cost. The codes corresponding to the paper are available on GitHub. Author Affiliation: School of Business and Economics, Humboldt University of Berlin, Unter den Linden 6, Berlin 10099, Germany * Corresponding author. Article History: Received 3 December 2020; Accepted 11 June 2021 Byline: Nikita Kozodoi [nikita.kozodoi@hu-berlin.de] (*,a), Johannes Jacob (a), Stefan Lessmann (a)

Keywords Combinatorial optimization; Clustering; Mixed integer programming; Branch-and-price Highlights * Branch-and-price algorithm for the Graph-Connected Clique-Partitioning Problem. * New heuristic algorithms for the same problem. * Complete computational experience for a large battery of instances. Abstract This paper studies the Graph-Connected Clique-Partitioning Problem (GCCP), a clustering optimization model in which units are characterized by both individual and relational data. This problem, introduced by Benati, Puerto, and Rodríguez-Chía (2017) under the name of Connected Partitioning Problem, shows that the combination of the two data types improves the clustering quality in comparison with other methodologies. Nevertheless, the resulting optimization problem is difficult to solve; only small-sized instances can be solved exactly, large-sized instances require the application of heuristic algorithms. In this paper we improve the exact and the heuristic algorithms previously proposed. Here, we provide a new Integer Linear Programming (ILP) formulation, that solves larger instances, but at the cost of using an exponential number of variables. In order to limit the number of variables necessary to calculate the optimum, the new ILP formulation is solved implementing a branch-and-price (B&P) algorithm. The resulting pricing problem is itself a new combinatorial model: the Maximum-weighted Graph-Connected Single-Clique problem (MGCSC), that we solve testing various Mixed Integer Linear Programming (MILP) formulations and proposing a new fast 'random shrink' heuristic. In this way, we are able to improve the previous algorithms: The B&P method outperforms the computational times of the previous MILP algorithms and the new random shrink heuristic, when applied to GCCP, is both faster and more accurate than the previous heuristic methods. Moreover, the combination of column generation and random shrink is itself a new MILP-relaxed matheuristic that can be applied to large instances too. Its main advantage is that all heuristic local optima are combined together in a restricted MILP, consisting in the application of the exact B&P method but solving heuristically the pricing problem. Author Affiliation: (a) Dipartimento di Sociologia e Ricerca Sociale, Università di Trento, Italy (b) IMUS and Departamento de Estadística e Investigación Operativa, Universidad de Sevilla, Spain (c) Departamento de Estadística e Investigación Operativa, Universidad de Cádiz, Spain * Corresponding author. Article History: Received 28 May 2020; Revised 24 May 2021; Accepted 25 May 2021 Byline: Stefano Benati [stefano.benati@unitn.it] (a), Diego Ponce [dponce@us.es] (b), Justo Puerto [puerto@us.es] (*,b), Antonio M. Rodríguez-Chía [antonio.rodriguezchia@uca.es] (c)

Keywords Multiple criteria analysis; Electre methods; Scoring methods; Non-compensatory composite indicators; Healthcare system evaluation Highlights * An outranking method for assigning a score to considered actions is proposed. * Actions are compared by means of an outranking with reference sets of actions. * A deck of the cards method is used to assign a value to reference sets of actions. * A range for its score is assigned to each action. * Conditions ensuring desirable properties for the proposed method are discussed. Abstract In this paper, we present (to the best of our knowledge) the first outranking method to assign a score to each alternative. This method is part of the Electre family, and we will call it Electre-Score. In contrast to the Multi-Attribute Value Theory methods, Electre-Score does not construct a value function for each criterion to proceed then to the aggregation into a single value. Instead, Electre-Score makes use of the outranking relations to make a comparison with sets of reference actions (to which a score is assigned) and assigns a score range to each action and a representative score in the range to each action. This is a more robust way of proceeding given the fragility of a single score. The reference actions have similar nature and characteristics of limiting profiles in Electre Tri-nB and the reference scores are assigned to them through the application of the deck of cards technique. Because this method uses outranking relations, it makes also possible to take into account the imperfect knowledge of data and to avoid systematic compensatory effects. Some fundamental theoretical results guaranteeing the consistency of the method, an illustrative example and a case study in the domain of healthcare system evaluation are also provided in this paper. Author Affiliation: (a) Department of Economics and Business, University of Catania, Catania, Italy (b) Université Paris-Dauphine, PSL Research University, CNRS (UMR 7243), LAMSADE, Paris, France (c) CEG-IST, Instituto Superior Técnico, Universidade de Lisboa, Av. Rovisco Pais 1, 1049-001, Lisbon, Portugal (d) Portsmouth Business School, Centre of Operations Research and Logistics (CORL), University of Portsmouth, Portsmouth, United Kingdom * Corresponding author. Article History: Received 22 July 2019; Accepted 13 May 2021 Byline: José Rui Figueira [figueira@tecnico.ulisboa.pt] (*,c), Salvatore Greco (a,d), Bernard Roy (b)

Keywords OR in marketing; Choice rules; Choice-based conjoint analysis; Preference shares; IIA property Highlights * The paper deals with shares of choice predictions from choice-based conjoint data. * Monte Carlo and empirical studies are proposed to compare different choice rules. * The direct use of Hierarchical Bayes draws substantially reduces prediction errors. * The superiority of using draws is independent of the similarity of alternatives. * The findings strongly suggest the use of Hierarchical Bayes draws for predictions. Abstract The use of Hierarchical Bayes (HB) estimation techniques for choice-based conjoint (CBC) data offers the opportunity to directly use HB draws for preference simulations. This paper analyzes the use of HB draws for shares of choice predictions. Five different choice rules are compared: the first choice rule applied to HB draws, the logit choice rule applied to HB draws, the randomized first choice rule, the traditional first choice rule and the traditional logit choice rule. Each two different holdout choice scenarios are constructed containing one time two extremely similar and the other time very unique alternatives to assess how well the choice rules tolerate the IIA property in predicting choice shares. We present a Monte Carlo study to systematically explore shares of choice predictions based on the different choice rules and further verify whether our findings hold in empirical settings. The key finding of our Monte Carlo study is that using HB draws either combined with the first choice rule or the logit choice rule substantially improves choice share predictions when compared to the other choice rules, regardless of the type of holdout choice scenario. While the logit choice rule applied to HB draws performs a touch better for simulated data, the first choice rule applied to HB draws provides the best choice share predictions for each of the five empirical data sets. Using HB draws does not only provide the best predictive validity but, more importantly, it is theoretically correct when applying a Bayesian estimation approach to CBC data. Author Affiliation: (a) State Office for Statistics Lower Saxony, Göttinger Chaussee 76, 30453 Hannover, Germany (b) Department of Marketing, Clausthal University of Technology, 38678 Clausthal-Zellerfeld, Germany (c) bms marketing research + strategy, Landsberger Straße 487, 81241 München, Germany * Corresponding author. Article History: Received 3 December 2018; Accepted 31 May 2021 Byline: Maren Hein [maren.hein@statistik.niedersachsen.de] (a), Nils Goeken [nils.goeken@tu-clausthal.de] (b), Peter Kurz [p.kurz@bms-net.de] (c), Winfried J. Steiner [winfried.steiner@tu-clausthal.de] (b,*)

Keywords Travelling salesman; Vehicle routing problem; Split demand; Pickup-and-delivery; Bike shared system Highlights * When customers accept several visits, the total cost of the route can be reduced. * On some applications clients may be used as buffers to store temporarily product. * The total cost can even be reduced if the initial load of the vehicle is unfixed. * We introduce a new procedure to check the feasibility of a route. * This manuscript describes a branch-and-cut algorithm to solve difficult instances. Abstract This paper deals with the problem of designing a minimum-cost route for a capacitated vehicle moving a commodity between a set of customers, allowing two features uncommon in the pickup-and-delivery literature. One feature is that a customer accepts to be visited several times, i.e., splitting a customer demand is allowed. The other feature is that a customer may be used as an intermediate location to collect and deliver commodity temporarily. The problem is called Split-Delivery One-Commodity Pickup-and-Delivery Travelling Salesman Problem, and finds applications in bike sharing systems where a single vehicle moves bikes between bike stations of a city district during the night to set the network to an initial configuration. The paper proposes a new branch-and-cut algorithm to find optimal solutions. A master problem solves a relaxed Mixed Integer Programming model, i.e., a model allowing all feasible solutions and also some invalid ones. A subproblem checks the feasibility of the master solutions and generates valid cuts when they are infeasible. Computational results on benchmark instances demonstrate the good performance of the algorithm compared with others in the literature. In particular, it solves benchmark instances with 60 customers that were unsolved. Author Affiliation: DMEIO, Facultad de Ciencias, Universidad de La Laguna, 38200 La Laguna, Tenerife, Spain * Corresponding author. Article History: Received 30 December 2020; Accepted 24 May 2021 (footnote)[white star] This work has been partially supported by the research project MTM2015-63680-R (MINECO/FEDER), PID2019-104928RB-100 (MINECO/FEDER, UE) and ProID2017010132 (Gobierno de Canarias/FEDER). Byline: Hipólito Hernández-Pérez [hhperez@ull.es] (a), Juan-José Salazar-González [jjsalaza@ull.es] (*,a)

Keywords Manufacturing; Sustainable product design; Environmental taxation; Sustainability Highlights * Proposes the use of an environmental tax to enhance DfE. * Examine three forms of environmental taxes. * Uncover leveraging effect of marginal DfE allowance. * Findings are robust with four extended modeling analyses. Abstract Nowadays, the concept of Design for Environment (DfE) is getting more and more important and popular for sustainable operations. Observations from real-world practices and extant literature indicate that currently, companies do not have strong incentive to establish DfE. To this end, this paper proposes the use of an environmental tax to enhance DfE, which will contribute to reducing wastes from the origins, improving the environmental performance and maximizing the social welfare. In this paper, by building a stylized analytical model, we examine three forms of environmental taxes (linear tax, constant tax and zero tax) and evaluate how they affect the producer's optimal DfE level. We interestingly find that the 'constant tax' does not encourage sustainable product design and the 'zero tax' is even better than the 'constant tax'. On the contrary, the 'linear tax' can promote sustainable product design and can be regarded as the best. We further discover that the 'linear tax' can help balance the DfE level, the stakeholders' benefits and the social welfare performance. The leveraging effect of marginal DfE allowance in the 'linear tax' is surprisingly important and useful. Finally, we extend the basic model to cover four cases to check the robustness of the analytical results in the basic model and several important managerial insights for the development of sustainable product design are generated from the stakeholders and environment perspectives. Author Affiliation: (a) School of Economics and Management, Southeast University, Nanjing 210093, China (b) Department and Graduate Institute of Business Administration, College of Management, National Taiwan University, Roosevelt Road, Taipei 10617, Taiwan (c) Institute of Supply Chain Analytics, Dongbei University of Finance and Economics, Dalian, Liaoning, 116025, China (d) Business School, Nankai University, Tianjin, 300071, China * Corresponding author. Article History: Received 5 October 2020; Revised 16 March 2021; Accepted 27 April 2021 (footnote) The authors sincerely thank the editor and reviewers for their very helpful and important comments, especially during the very difficult time with COVID19 pandemic. They feel very grateful and hope all editors and reviewers stay safe from the virus. Byline: Ya-Jun Cai (a), Tsan-Ming Choi [tmjchoi@gmail.com] (b,*), Lipan Feng (c), Yongjian Li (d)

Keywords Pricing; Dynamic pricing; Product acquisition; Dynamic programming; Secondary market Highlights * This paper is motivated by firms that acquire and resell pre-owned items, such as cell phones. * In our model, the firm dynamically adjusts both selling and acquisition prices. * We show how the prices and the firm's margin change with time and inventory. * We devise heuristic policies with static prices based on a deterministic approximation. * We evaluate the value of dynamic pricing relative to static prices. Abstract We analyze a novel dynamic pricing problem, where not only the demand depends on the dynamic selling price, but also the supply can be influenced through the dynamic acquisition price. This paper is motivated by firms that acquire and resell pre-owned items, such as cell phones. A firm with fixed inventory is always under pressure to sell the remaining inventory, yet when acquisition opportunities arise, the firm may become under pressure to acquire inventory in anticipation of a surge in demand. We establish conditions under which the firm will find itself under pressure to acquire versus pressure to sell. We study how the optimal acquisition and selling prices, and the difference between them (the optimal margin) behaves with respect to time and inventory. In addition, we devise heuristics with fixed prices based on a deterministic approximation, and we evaluate how much the firm gains by using a dynamic acquisition or selling price as opposed to fixing them at the beginning of the horizon. Author Affiliation: (a) College of Business Administration, Kuwait University, Shuwaikh, 13015, Kuwait (b) School of Statistics, Southwestern University of Finance and Economics, Chengdu, 611130, China (c) Carey Business School, Johns Hopkins University, Baltimore, MD 21202, USA (d) Kelley School of Business, Indiana University, Bloomington, IN 47405, USA * Corresponding author. Article History: Received 6 August 2019; Accepted 28 April 2021 Byline: Mohammad Ghuloum [mghuloum@cba.edu.kw] (a), Boqian Song [songbq@swufe.edu.cn] (*,b), Goker Aydin [goker.aydin@jhu.edu] (c), Gilvan C. Souza [gsouza@indiana.edu] (d)

Keywords Forecasting; Elo rating; Tennis; Betting strategy Highlights * We propose an extension of the Elo rating system labelled Weighted Elo (WElo). * WElo includes the 'hot hand' phenomenon in the weighting function. * WElo approach outperforms the most popular forecasting methods in tennis. * According to a simple betting strategy WElo provides profitable opportunities. * We provide a R package called 'welo' to estimate Elo and WElo rating systems. Abstract Originally applied to tennis by the data journalists of FiveThirtyEight.com, the Elo rating method estimates the strength of each player based on her/his career as well as the outcome of the last match played. Together with the regression-based, point-based and paired-comparison approaches, the Elo rating is a popular method to predict the probability of winning tennis matches. Notwithstanding its widely recognized merits in terms of ease of reproducibility and good performance, the Elo method does not completely take into account the current form of each player and their recent performances. This paper proposes a new version of the Elo rating method, labelled Weighted Elo (WElo), where the standard Elo updating is additionally weighted according to the scoreline of the players' last match. The proposed method considers not only if a player has won (lost) a match, but also how the victory (defeat) was achieved. In the empirical application, the forecasting performance of the WElo method is evaluated and compared against the most popular forecasting methods in tennis, using a sample of over 60,000 men's and women's professional matches. Overall, the WElo method outperforms all these competing methods. Moreover, it provides meaningfully profitable opportunities, according to a simple betting strategy. Author Affiliation: (a) Department of Economics, University of Bologna, Italy (b) Department MEMOTEF, Sapienza University of Rome, Italy (c) Department of Economics, University of Bologna, Italy * Corresponding author. Article History: Received 20 September 2020; Revised 17 March 2021; Accepted 8 April 2021 (footnote)[white star] We are grateful to the Editor, Ruud Teunter, and three anonymous referees for their helpful comments on the earlier version of this paper. We also thank the participants at the 12th International Conference of the ERCIM on Computational and Methodological Statistics (CMStatistics) in London (December 2019) for their useful comments. Byline: Giovanni Angelini [g.angelini@unibo.it] (a), Vincenzo Candila [vincenzo.candila@uniroma1.it] (b), Luca De Angelis [deangelis@unibo.it] (*,c)

Keywords Flexible manufacturing systems; Scheduling; Iterated local search; Tool switching; Sequence-dependent setup times Highlights * The job sequencing and tool switching problem with non-identical parallel machines is studied. * Three different objective functions are considered. * Machine-dependent processing and tool switching times are taken into account. * A MILP model and several ILS algorithms are proposed. * Computational experiments show the efficiency of the algorithms. Abstract In this paper, a new generalization of the uniform job sequencing and tool switching problem is presented. It considers non-identical parallel machines, which differ in tool magazine capacity and setup times. Applications of the problem can be found in the metal working or semiconductor manufacturing industries when the jobs require different tool sets for processing. The paper provides a Mixed Integer Programming formulation for the job sequencing and tool switching problem with machine-dependent processing and tool switching times under the consideration of three conflicting objectives (minimizing the total flowtime, minimizing the makespan and minimizing the total number of tool switches). Different efficient construction heuristics and Iterated Local Search methods are proposed and evaluated, using 640 new and publicly available instances. The results of the construction heuristics, the Iterated Local Search schemes and a Mixed Integer Programming Solver are discussed. The extensive computational experiments show the merit of the perturbation strategy in order to overcome local optima. It is shown that certain methods are more or less suitable depending on the objective function. Author Affiliation: School of Business, Economics and Information Systems, Chair of Information Systems (Information and IT Service Management), University of Passau, Innstrasse 43, 94032 Passau, Germany Article History: Received 27 March 2020; Accepted 9 May 2021 (footnote)1 http://www.wi.uni-passau.de/ Byline: Dorothea Calmels [dorothea.calmels@uni-passau.de] (1)

Keywords OR in scientometrics; Research evaluation of institutions; Departmental h -index dynamics; Forecast of research assessment Highlights * We propose a new approach to study the research evaluation of departments. * We use linear and exponential models for the dynamics of the departmental h-index. * We investigate the effects of the structure by age of the faculty on the h-index. * We present two models that better fit the results of research evaluation exercises. * The models are tested on data of UK departments. Abstract This paper proposes a new approach to study the outcomes of the research evaluation of departmental structures and to predict the results of the next research evaluation exercise. To this aim we use the information provided by the dynamics of the departmental h-index; hence, in the first part of the paper we propose two models for these dynamics: a linear and an exponential model. Moreover, we investigate its determinants, especially the influence of the structure by age of the department faculty. Afterwards, we build two new models that improve the goodness of fit for the outcomes of the research assessment: a linear model stemming from the assumption of linear dynamics for the departmental h-index, and a log-linear model consistent with the assumption of an exponential dynamics. These models are tested on data of the UK departments of four different scientific fields (Biology, Chemistry, Physics and Sociology) and the results show that this approach can be successfully applied. Author Affiliation: (a) Department of Economics, Ca' Foscari University of Venice, Cannaregio 873, Venezia 30121, Italy (b) Department of Economics, Ca' Foscari University of Venice, Cannaregio 873, Venezia 30121, Italy * Corresponding author. Article History: Received 10 August 2020; Accepted 8 May 2021 Byline: Antonella Basso [basso@unive.it] (*,a), Giacomo di Tollo [giacomo.ditollo@unive.it] (b)

Keywords Finance; Optimal liquidation problem; Illiquid market; Markov-modulated poisson process; Hawkes processes Highlights * Proposing an analytical solution for the optimal liquidation problem in an illiquid market. * Applying trading boundaries built-in on the market depth to adjust the speed of trading. * Expressing the distribution of the limit order book analytically. * Modelling dynamics of incoming orders and measuring the price impact using Hawkes Processes. * Investigating the dependency of optimal trading strategies on market characteristics. Abstract In this research, we develop a trading strategy for the optimal liquidation problem of large-order trading, with different market microstructures, in an illiquid market. We formulate the liquidation problem as a discrete-time Markov decision process. In this market, the flow of liquidity events can be viewed as a point process with stochastic intensity. Based on this fact, we model the price impact as a linear function of a self-exciting dynamic process. Our trading algorithm is designed in such a way that when no favourite orders arrive in the Limit Order Book (LOB), the optimal solution takes offers from the lower levels of the LOB. This solution might contradict conventional optimal execution methods, which only trade with the best available limit orders; however, our findings show that the proposed strategy may reduce final inventory costs by preventing orders not being filled at earlier trading times. Furthermore, the results indicate that an optimal trading strategy is dependent on characteristics of the market microstructure. Author Affiliation: (a) Department of Finance and Accounting, Rennes School of Business, 2 Rue Robert d'Arbrissel, Rennes 35065, France (b) Department of Probability and Mathematical Statistics, Charles University, Sokolovska 83, 18675, Praha 8, Czech Republic (c) Department of Finance, Frankfurt School of Finance and Management, Adickesalle 32-34 60322, Frankfurt am Main, Germany * Corresponding author. Article History: Received 18 November 2020; Accepted 17 May 2021 (footnote)[white star] We thank two anonymous referees, and the Editor, Emanuele Borgonovo, for helpful comments that significantly improved the paper. We are grateful for valuable feedback from Stefan Kassberger and Falko Fecht. The former title of the paper is 'Optimum Liquidation Problem Associated with the Poisson Cluster Process'. Byline: Amirhossein Sadoghi [amirhossein.sadoghi@rennes-sb.com] (*,a), Jan Vecer [vecer@karlin.mff.cuni.cz] (b,c)

Keywords Humanitarian logistics; Cash-based interventions; Syrian refugee crisis; Location-routing Highlights * Cash-based interventions (CBI) play an important role in aiding people post-crisis. * There is limited research supporting the roll-out of cash-based interventions. * CBI refugee coverage is optimized using bi-objective mathematical programming. * CBI delivery capacity will be underutilized if mobile facilities are not deployed. * Having few facilities open for a long period is better than many for a short period. Abstract This paper addresses a humanitarian logistics problem connected with the Syrian refugee crisis. The ongoing conflict in Syria has caused displacement of millions of people. Cash-based interventions play an important role in aiding people in the post-crisis period to enhance their well-being in the medium and longer term. The paper presents a study on how to design a network of administrative facilities to support the roll-out of cash-based interventions. The resulting multi-level network consists of a central registration facility, local temporary facilities, mobile facilities and vehicles for door-to-door visits. The goal is to reach the maximum number of eligible beneficiaries within a specified time period while minimizing logistics costs, subject to a limit on total security risk exposure. A mixed integer programming model is formulated to optimize the inter-related facility location and routing decisions under multiple objectives. The authors develop a hierarchical multi-objective metaheuristic algorithm to obtain efficient solutions. An application of the model and the solution algorithm to real data from a region in the southeast of Turkey is presented, with associated managerial insights. Author Affiliation: (a) Nottingham Business School, Nottingham Trent University, Nottingham NG1 4FQ, UK (b) School of Management, University of Bath, BA1 7AY Bath, UK (c) Operations Research and Logistics group, Wageningen University, Hollandseweg 1,Wageningen, The Netherlands (d) College of Engineering, Koç University, Sariyer Istanbul 34450, Turkey (e) Wolfson School of Mechanical, Electrical and Manufacturing Engineering, Loughborough University, Leicestershire LE11 3TU, UK (f) Department of Industrial Engineering, Bilkent University, Bilkent 06800 Ankara, Turkey (g) Department of Logistics Management, Business School, Izmir University of Economics, Balçova 35330 Izmir, Turkey * Corresponding author. Article History: Received 13 February 2020; Accepted 28 April 2021 Byline: Ramez Kian [ramez.kian@ntu.ac.uk] (a), Günes Erdogan [g.erdogan@bath.ac.uk] (b), Sander de Leeuw [sander.deleeuw@wur.nl] (*,a,c), F. Sibel Salman [ssalman@ku.edu.tr] (d), Ehsan Sabet [E.Sabet@lboro.ac.uk] (e), Bahar Y. Kara [bkara@bilkent.edu.tr] (f), Muhittin H. Demir [muhittin.demir@ieu.edu.tr] (g)

Keywords Finance; Trade credit; Vertical integration; Supply chain coordination; Real options Highlights * We analyze trade credit within a real options framework. * We show that trade credit serves as a coordination device. * We show that the supplier can optimally choose to offer trade credit for free. * We study optimal trade credit design. * Regulation on trade credit maturity leads to a decrease in trade credit. Abstract This paper provides a theoretical analysis of trade credit within a real options framework. We show that under trade credit the buyer delays the decision to stop production, getting closer to the supply chain optimal stopping decision. Therefore, trade credit may serve as a coordination device. The supplier can optimally choose to offer trade credit for free, since this will guarantee her business for a longer period of time. Optimal trade credit design is analyzed for an integrated supply chain (cooperative solution) and for external procurement (Nash bargaining and Stackelberg solutions). When regulation imposes a limit on trade credit maturity, the wholesale price is reduced, trade credit decreases and internal procurement increases. The model's predictions are in line with recent empirical evidence on the effects of regulation in the retail industry. Author Affiliation: (a) Universitat Autonoma de Barcelona, Campus de Bellatera, Barcelona, 08193, Spain (b) Univ Rennes, CNRS, CREM - UMR6211, Rennes F-35000, France * Corresponding author. Article History: Received 30 December 2020; Accepted 20 April 2021 (footnote)[white star] We are grateful to Emanuele Borgonovo (the editor-in-chief) and an anonymous referee for constructive comments that helped us improve the paper. We thank Pauline Gandré, Pascal François, Hidekazu Yoshioka, Nicos Koussis, as well as conference participants at the AFFI conference, Québec 2019 and the Real Options conference, London 2019, for valuable comments and suggestions. Florina Silaghi, Serra Húnter Fellow, gratefully acknowledges financial support from the Ramón Areces Foundation, from the Spanish National Research Agency through the Juan de la Cierva Incorporation Grant IJCI-2017-31641 and from the Spanish Ministry of Economics and Competitiveness through project ECO2017-86054-C3-1-R. Franck Moraux gratefully acknowledges the financial support of the French National Research Agency through the project FILEAS-FOG (ANR-17-CE10-0001-01). Byline: Florina Silaghi [florina.silaghi@uab.cat] (*,a), Franck Moraux [franck.moraux@univ-rennes1.fr] (b)

Keywords Project management; Dependence modelling; Large-scale risks; Simulation Highlights * Present a multi-factor dependence model for large-scale project risk assessment. * Elicit structured association (SA) factors from ubiquitous project plans. * Transform SA factors into a mathematically consistent correlation matrix. * Evaluate performance against distribution types, skewness, and project sizes. * Demonstrate scalability to large-scale projects given partially specified dependence. Abstract This paper examines the high-dimensional dependence modelling problem in the context of project risk assessment. As the dimension of uncertain performance units (i.e., itemized costs and activity times) in a project increases, specifying a feasible correlation matrix and eliciting relevant pair-wise information, either from historical data or with expert judgement, becomes practically unattainable or simply not economical. This paper presents a factor-driven dependence elicitation and modelling framework with scalability to large-scale project risks. The multi-factor association model (MFAM) accounts for hierarchical relationships of multiple association factors and provides a closed-form solution to a complete and mathematically consistent correlation matrix. Augmented with the structured association (SA) technique for systematic identification of hierarchical association factors, the MFAM offers additional flexibility of utilizing the minimum information available in standardized, ubiquitous project plans (e.g., work breakdown structure, resource allocation, or risk register), while preserving the computational efficiency and the scalability to high dimensional project risks. Numerical applications and simulation experiments show that the MFAM, further combined with extended analytics (i.e., parameter calibration and optimization), provides credible risk assessments (with accuracy comparable to full-scale simulation) and further enhances the realism of dealing with high-dimensional project risks utilizing all relevant information. Author Affiliation: Project and Supply Chain Management, Black School of Business, Penn State Erie, the Behrend College, 5101 Jordan Road, Erie 16563-1400, PA, United States Article History: Received 8 July 2020; Accepted 23 April 2021 Byline: Byung-Cheol Kim [buk70@psu.edu]

Keywords Heuristics; Lot sizing; Scheduling; Non-triangular setup; Matheuristic Highlights * Local search strategies with path-relinking and kernel search to an integrated problem. * Matheuristics specially designed to approach non-triangular sequence-dependent setups. * Methods significantly outperformed a solver in most of the tested instances. Abstract This paper approaches the integrated lot sizing and scheduling problem (ILSSP), in which non-identical machines work in parallel with non-triangular sequence-dependent setup costs and times, setup carry-over and capacity limitation. The aim of the studied ILSSP, here called ILSSP-NT on parallel machines, is to determine a production planning and tasks sequencing that meet period demands without delay and in such a way that the total costs of production, machine setup and inventory are minimized. The dearth of literature on the ILSSP-NT, despite the increasing amount of applications in the industrial sector, mainly in the food processing industry, motivated us to conduct this study. In this paper, we propose efficient methods to solve the ILSSP-NT on parallel machines. The methods virtually consist in the hybridization of the relax-and-fix and fix-and-optimize methods with the path-relinking and kernel search heuristics. To assess how well the heuristics solve the ILSSP-NT on parallel machines, we compared their results with those of the CPLEX solver with a fixed CPU time limit. The proposed matheuristics significantly outperformed CPLEX in most of the tested instances. Author Affiliation: Instituto de Ciência e Tecnologia, Universidade Federal de São Paulo -- UNIFESP, Av. Cesare G. Lattes, 1201 São José dos Campos, SP, Brazil * Corresponding author. Article History: Received 17 April 2019; Accepted 27 March 2021 Byline: Desiree M. Carvalho [dmcarvalho@unifesp.br] (*), Mariá C.V. Nascimento [mcv.nascimento@unifesp.br]

Keywords Scheduling; Multiobjective optimization; Flexible job-shop scheduling with batching; Metaheuristics; OR in semiconductor manufacturing Highlights * A flexible job-shop scheduling problem with complex constraints is studied. * Multiple industrially relevant criteria are optimized using Simulated Annealing. * A lexicographic order and weights are both used to model users' preferences. * A novel criterion modeling the satisfaction of production targets is introduced. * Numerical results on large industrial instances are discussed. Abstract In this paper, we are concerned with the resolution of a multiobjective complex job-shop scheduling problem stemming from semiconductor manufacturing. To produce feasible and industrially meaningful schedules, this paper extends the recently proposed batch-oblivious approach by considering unavailability periods and minimum time lags and by simultaneously optimizing multiple criteria that are relevant in the industrial context. A novel criterion on the satisfaction of production targets decided at a higher level is also proposed. Because the solution approach must be embedded in a real-time application, decision makers must express their preferences before the optimization phase. In addition, a preference model is introduced where trade-off is only allowed between some criteria. Two a priori multiobjective extensions of Simulated Annealing are proposed, which differ in how the simultaneous use of a lexicographic order and weights is handled when evaluating the fitness. A known a posteriori approach of the literature is used as a benchmark. All the metaheuristics are embedded in a Greedy Randomized Adaptive Search Procedure. The different versions of the archived GRASP approach are compared using large industrial instances. The numerical results show that the proposed approach provides good solutions regarding the preferences. Finally, the comparison of the optimized schedules with the actual factory schedules shows the significant improvements that our approach can bring. Author Affiliation: (a) Department of Accounting and Operations Management, BI Norwegian Business School, Oslo, 0484, Norway (b) Mines Saint-Etienne, Univ Clermont Auvergne, CNRS, UMR 6158 LIMOS, CMP, Department of Manufacturing Sciences and Logistics, Gardanne 13541, France (c) Untis GmbH, 2000 Stockerau, Austria (d) Planimize, Paris, France * Corresponding author at: BI Norwegian Business School, Department of Accounting, Auditing and Business Analytics, Nydalsveien 37, N-0484 Oslo Norway. Article History: Received 23 October 2019; Accepted 17 March 2021 Byline: Karim Tamssaouet [karim.tamssaouet@bi.no] (*,a,b), Stéphane Dauzère-Pérès [dauzere-peres@emse.fr] (a,b), Sebastian Knopp [mail@sebastian-knopp.de] (c), Abdoul Bitar [abdlbitar@gmail.com] (d), Claude Yugma [yugma@emse.fr] (b)

Keywords Portfolio optimisation; Behavioural finance; Cumulative prospect theory; Investor memory; Naïve investment strategy Highlights * We investigate the performance of behavioural investment strategies. * We propose a new behavioural investment objective function. * We compare the behavioural portfolio with rational and naïve portfolios. * We use both EUT-based and CPT-based evaluation criteria. Abstract In this paper, we investigate the performance of behavioural portfolio strategies. We incorporate the short-term and long-term memory of the investor, thus recasting the behavioural portfolio choice process in a dynamic setting. We evaluate the out-of-sample performance of a behavioural investor in relation to both a naïve investor who invests in an equally weighted portfolio and a rational investor, who maximises expected mean-variance utility. We report a number of findings. First, from an expected utility perspective, neither the rational investor nor the CPT investor achieves a risk-adjusted return or certainty equivalent return that significantly outperforms that of the naïve investor. Second, from a CPT utility perspective, the behavioural investor outperforms both the rational and naïve investors. Third, the CPT investor typically displays highly concentrated, lottery-like asset allocations, low turnover and highly stable portfolio allocations. Fourth, the addition of the investor's memory into the portfolio choice process increases both diversification and turnover, leading to improved investment performance. Finally, by allocating more weight to positively skewed assets and increasing portfolio concentration, the probability weighting function has more impact than the utility function on the behavioural investor's performance. Our results are robust to the choice of reference return, estimation sample size, probability estimates, the probability weighting function and portfolio weight constraints. Author Affiliation: (a) School of Accounting and Finance, University of Bristol, 15-19 Tyndalls Park Road, Bristol, BS8 1PQ, UK (b) University of Dundee School of Business, 3 Perth Road, Nethergate, Dundee DD1 4HN, UK * Corresponding author. Article History: Received 8 October 2019; Accepted 23 April 2021 (footnote) Parts of the paper were written while Mazibas was the Appleby Research Fellow at the University of Exeter. We are grateful to Howard Appleby for his generous support. We would like to thank the editor and four reviewers for their constructive comments and suggestions. Byline: Richard D. F. Harris [Richard.Harris@bristol.ac.uk] (a,*), Murat Mazibas [M.Mazibas@dundee.ac.uk] (b)

Keywords Combinatorial optimization; Facility location problem; Routing; VNS; Waste collection Highlights * We model the introduction of an underground waste container system. * We develop different solution approaches to tackle this problem. * Integrated solution approaches do not necessarily lead to cost reduction. * The introduction of an underground waste container system indicates saving potential. Abstract In Germany, household waste is collected via a door-to-door collection system. This form of waste collection is already reaching its limits due to the increasing volume of household waste and lack of space for even more dustbins, especially in densely populated cities. In this paper we investigate an alternative collection system: an underground waste collection system in which waste would be collected at central locations in so-called underground containers with a much larger capacity. If an underground container system is to be introduced to replace the present door-to-door waste collection system, a number of decisions of varying scope have to be made: Where should the underground containers be placed? How much capacity should be provided at which collection site? Which household should dispose of its waste where? How often, when and in what order should the waste collection vehicles empty the containers? In this paper we develop a model to illustrate the combination of the underlying problems by extending a problem known from literature. We present both hierarchical and integrated approaches to solve it. The proposed solution approaches combine in different ways two variable neighborhood search heuristics and a mixed integer programming-based exact method. We show that great benefits can be achieved through the use of integrated solution approaches. But we show too that integrated solution approaches do not necessarily dominate hierarchical ones. Author Affiliation: Department of Business Administration and Economics, Bielefeld University, Bielefeld, Germany * Corresponding author. Article History: Received 2 March 2020; Accepted 27 February 2021 Byline: Sina Gläser [sina.glaeser@uni-bielefeld.de] (*), Mareike Stücken

Keywords Reliability; Condition-based maintenance; Markov decision process; Finite horizon; Bivariate gamma process Highlights * We develop a condition-based maintenance model in a finite-time horizon. * We consider a system with two heterogeneous dependent components. * We determine the boundaries for various maintenance actions. * We investigate the influence of economic dependence on optimal maintenance policy. * We investigate the effect of dependency between the degradation processes. Abstract This paper analyzes a condition-based maintenance (CBM) model for a system with two heterogeneous components in which degradation follows a bivariate gamma process. Unlike the traditional CBM formulation that assumes an infinite planning horizon, this paper evaluates the maintenance cost in a finite planning horizon, which is the practical case for most systems. In the proposed CBM policy, both components are periodically inspected and a preventive or corrective replacement might be carried out based on the state of degradation at inspection. The CBM model is formulated as a Markov decision process (MDP) and dynamic programming is used to compute the expected maintenance cost over a finite planning horizon. The expected maintenance cost is minimized with respect to the preventive replacement thresholds for the two components. Unlike an infinite-horizon CBM problem, which leads to a stationary maintenance policy, the optimal policy in the finite-horizon case turns out to be non-stationary in the sense that the optimal actions vary at each inspection epoch. A numerical example is presented to illustrate the proposed model and investigate the influence of economic dependency and correlation between the degradation processes on the optimal maintenance policy. Numerical results show that a higher dependence between the degradation processes actually reduces the maintenance cost, while a higher economic dependence leads to higher preventive replacement thresholds. Author Affiliation: (a) Department of Management Science, University of Strathclyde, Glasgow G1 1XQ, UK (b) Department of Civil and Environmental Engineering, University of Waterloo, Waterloo, ON N2L 3G1, Canada (c) Department of Logistics and Maritime Studies, The Hong Kong Polytechnic University, Kowloon, Hong Kong (d) College of Management and Economics, Tianjin University, Tianjin, China * Corresponding author. Article History: Received 27 August 2019; Accepted 5 March 2021 Byline: Bin Liu (a), Mahesh D. Pandey (b), Xiaolin Wang (c), Xiujie Zhao [xiujiezhao@tju.edu.cn] (*,d)

Keywords Data envelopment analysis; Cost efficiency; Uncertainty; Robust optimisation Highlights * Assessing cost efficiency measures under endogenous and exogenous uncertainties. * Presenting robust DEA models to tackle uncertain inputs and outputs. * Developing a pair of robust models to attend to uncertain input prices. * Allowing the decision maker to control the uncertainty level. * Comparing the developed models with other existing approaches. Abstract Measuring economic and cost efficiency receives ever-increasing attention of the executives and managers of small-and medium-sized enterprises (SMEs) to minimise total production costs. The conventional Farrell cost efficiency (CE) as a key determinant requires the precise information on inputs, outputs and input prices, while in praxis uncertainty is inherent and inevitable in data and its negligence conceivably results in a dire approximation for CE measures. This paper is concerned with Farrell CE in situations of both endogenous and exogenous uncertainty. The source of uncertainty allows us to define two different scenarios; (i) in situations of endogenous uncertainty in input and output data where the uncertainty is affected by the decision maker's decisions, and (ii) in situations of uncertain prices for inputs where the uncertainty is exogenously given. In the first scenario, the theory of robust optimisation is adopted to develop the robust data envelopment analysis (DEA) models with the aim of grappling uncertainties in input and output data when measuring technical and cost efficiencies. The second scenario aims to accommodate uncertainties on price information by developing a pair of robust DEA models based upon robust optimisation estimating the upper and lower bounds for CE measures. This unprecedented study helps us to provide a generalised framework for economic efficiency with uncertainties in which conventional properties of Farrell measures are fulfilled. In addition to comparing the developed approach in this paper with other existing approaches through a simple numerical example, the usefulness and applicability of the suggested framework are minutely studied in an empirical application in the context of allocation problems. Author Affiliation: (a) Department of Management and Entrepreneurship, Leicester Castle Business School, De Montfort University, Hugh Aston Building, The Gateway, Leicester LE1 9BH, UK (b) Department of Mathematics, Faculty of Sciences, Golestan University, Gorgan, Iran * Corresponding author. Article History: Received 2 June 2020; Revised 15 December 2020; Accepted 9 March 2021 Byline: Adel Hatami-Marbini [adel.hatamimarbini@dmu.ac.uk] (a,*), Aliasghar Arabmaldar [n.arab@pgs.usb.ac.ir] (b)

Keywords Queueing; Non-cooperative games; Measure of efficiency; Price of anarchy; Braess paradox, Highlights * The paper surveys the literature on efficiency loss in stochastic queueing systems. * The price of anarchy along with other measures of efficiency are discussed. * The queueing models are classified and briefly described. Abstract This paper summarizes the literature on efficiency loss when agents selfishly optimize their utility in stochastic queueing systems. The price of anarchy is the most popular measure for quantifying this loss, but we also discuss other measures. The queueing models are introduced briefly and results on the loss of efficiency are discussed. Author Affiliation: (a) Department of Basic Science and Humanities, University of Engineering and Management Kolkata, Kolkata 700160, India (b) Department of Statistics and Operations Research, Tel Aviv University, Tel Aviv 69978, Israel * Corresponding author at: The research was done while the corresponding author was on a postdoctoral fellowship at the Department of Statistics and Operations Research, Tel Aviv University. 69978 Tel Aviv, Israel. Article History: Received 13 August 2020; Accepted 31 March 2021 (footnote)[white star] This research was supported by the Israel Science Foundation (Grant No. 1571/19). Byline: Souvik Ghosh [souvikghosh589@gmail.com] (*,b,a), Refael Hassin [hassin@tauex.tau.ac.il] (b)

Keywords Cutting; Steel industry; Mixed integer linear optimisation Highlights * A Mixed Integer Linear Optimization Model is proposed for a 1.5D Slitting Problem. * The leftovers may be used in the future, otherwise they are considered as scrap. * The model has been validated using real-world data from a Spanish company. * The model is solved in only a few minutes, while the company needs several hours. * With the model users improve order accuracy and drastically reduce usable leftovers. Abstract From an economic point of view, the steel industry plays an important role and, when it comes to responding to new challenges, innovation is a crucial factor. This paper proposes a mathematical methodology to solve the slitting problem in a steel company located in Europe. The slitting problem occurs when large width steel coils are slit into narrower coils, known as strips, to meet the requirements of the customers. A major challenge here is defining a slitting plan to fulfil all these requirements, as well as ongoing operational constraints and customer demands. The company looks for a reduction of the leftovers generated in the entire process, while maximising the overall accuracy of the orders. These leftovers may be used in the future as part of new orders provided they are able to respond to specific requirements, or otherwise they are discarded and considered as scrap. This paper introduces a novel mixed integer linear optimisation model to respond to a specific slitting problem. The model is validated with real data and it outperforms the results obtained by the company in different ways: by adjusting the orders that are to be served, by reducing the amount of scrap and by using the retails for future orders. Furthermore, the model is solved in only a few minutes, while the company needs several hours to prepare the scheduling in the current operating process. Author Affiliation: (a) IDOM Consulting, Engineering, Architecture, Spain (b) Department of Computer Science, Computer Architecture, Computer Languages & Information Systems, Statistics & Operations Research, Universidad Rey Juan Carlos, Spain (c) Departamento de Estadística e Investigación Operativa, Instituto de Matemática Interdisciplinar, Universidad Complutense de Madrid, Spain * Corresponding author at: IDOM Consulting, Engineering, Architecture, Spain. Article History: Received 12 August 2020; Accepted 22 February 2021 (footnote)[white star] This work was supported by the Government of the Region of Madrid (M. Sierra-Paradinas and O. Soto-Sánchez, grant IND2018/TIC-9614) and Government of Spain (A. Alonso-Ayuso and F.J. Martín-Campo, grant RTI2018-094269-B-I00). Byline: María Sierra-Paradinas [maria.sierrap@urjc.es] (*,a,b), Óscar Soto-Sánchez (b), Antonio Alonso-Ayuso (b), F. Javier Martín-Campo (c), Micael Gallego (b)

Keywords Scheduling; Open shop; Cooperative game theory; Core allocation rule; Balancedness Highlights * We take a game theoretical approach to open shop scheduling problems. * Unit open shop scheduling games need not be neither convex nor [sigma]-component additive. * We provide an allocation rule that lies in the unit open shop games. * Balancedness cannot be extended to general open shop scheduling games. Abstract This paper takes a game theoretical approach to open shop scheduling problems to minimize the sum of completion times. We assume that there is an initial schedule to process the jobs (consisting of a number of operations) on the machines and that each job is owned by a different player. Thus, we can associate a cooperative TU-game to any open shop scheduling problem, assigning to each coalition the maximal cost savings it can obtain through admissible rearrangements of jobs' operations. A number of different approaches to admissible schedules for a coalition are introduced and, in the main result of the paper, a core allocation rule is provided for games arising from unit (execution times and weights) open shop scheduling problems for the most of these approaches. To sharpen the bounds of the set of open shop scheduling problems that result in games that are balanced, we provide two counterexamples: one for general open shop problems and another for further relaxations of the definition of admissible rearrangements for a coalition. Author Affiliation: (a) Departamento de Análisis Económico, Universidad Nacional de Educación a Distancia (UNED), Madrid, Spain (b) Departament de Matemàtica Econòmica, Financera i Actuarial, and BEAT, Universitat de Barcelona, Barcelona, Spain (c) Universitat Oberta de Catalunya, Barcelona, Spain * Corresponding author. Article History: Received 30 July 2019; Accepted 10 February 2021 Byline: Ata Atay [aatay@cee.uned.es] (*,a), Pedro Calleja [calleja@ub.edu] (b), Sergio Soteras [ssoterass@uoc.com] (c)

Keywords Combinatorial optimization; Capacitated vehicle routing problem (CVRP); Adaptive iterated local search (ILS); Path-ReLinking Highlights * Diversity in Iterated Local Search (ILS) by perturbation and acceptance parameters. * Diversity control is addressed with the proposed adaptive ILS (AILS). * AILS with Path-Relinking (AILS-PR) enabled a good diversification and intensification. * AILS-PR outperformed state-state-of-the-art CVRP metaheuristics. Abstract Metaheuristics are widely employed to solve hard optimization problems, like vehicle routing problems (VRP), for which exact solution methods are impractical. In particular, local search-based metaheuristics have been successfully applied to the capacitated VRP (CVRP). The CVRP aims at defining the minimum-cost delivery routes for a given set of identical vehicles since each vehicle only travels one route and there is a single (central) depot. The best metaheuristics to the CVRP avoid getting stuck in local optima by embedding specific hill-climbing mechanisms such as diversification strategies into the solution methods. This paper introduces a hybridization of a novel adaptive version of Iterated Local Search with Path-Relinking (AILS-PR) to the CVRP. The major contribution of this paper is an automatic mechanism to control the diversity step of the metaheuristic to allow it to escape from local optima. The results of experiments with 100 benchmark CVPR instances show that AILS-PR outperformed the state-of-the-art CVRP metaheuristics. Author Affiliation: Instituto de Ciência e Tecnologia, Universidade Federal de São Paulo (UNIFESP), Av. Cesare M. G. Lattes, 1201, Eugênio de Mello, São José dos Campos, SP CEP 12247-014, Brasil * Corresponding author. Article History: Received 27 April 2020; Accepted 5 February 2021 Byline: Vinícius R. Máximo, Mariá C.V. Nascimento [mcv.nascimento@unifesp.br] (*)

Keywords Data envelopment analysis; Olympic; Multiplier formulation; Decomposition of efficiency; Prediction of medal tallies Highlights * The paper proposes a model with restricted multipliers for Olympic evaluation. * The paper provides a decomposition of the efficiency of Olympic achievements. * The paper examines Brazil's target for the 2016 Summer Olympic Games. * The paper also examines Japan's target for the upcoming Summer Olympic Games. Abstract Data envelopment analysis (DEA) is a useful tool for measuring the relative efficiencies of participating nations in the Olympic Games. DEA models with restricted multipliers have been used to refine efficiency evaluations by imposing additional information. Existing DEA models for evaluating Olympic medals do not focus on multiplier restrictions regarding input. To fill this research gap, this study incorporates a data fitting technique of medal prediction using ordinary least squares regression in input multiplier restrictions of the conventional DEA model. We show that the efficiency of the proposed model can be decomposed into the achievement ratio of substantial medal total and the unit value index of medals. Such decompositions can be used to analyze the effectiveness of host nations and athlete development initiatives. For an illustrative empirical application, we examine the target that the Brazilian Olympic Committee (BOC) set for the 2016 Summer Olympic Games (Rio 2016). Our results explain the extremely high feasibility of Brazil's target of being in the top 10 medals table in Rio 2016. Author Affiliation: Tokyo University of Science, 1-3 Kagurazaka, Shinjuku-ku, Tokyo 162-8601, Japan * Corresponding author. Article History: Received 5 October 2020; Accepted 19 February 2021 Byline: Kazuyuki Sekitani [sekitani@rs.tus.ac.jp], Yu Zhao [yu.zhao@rs.tus.ac.jp] (*)

Keywords Scheduling; Rail transportation; Rapid-shunting facilities; Platooning; Train scheduling Highlights * This paper investigates the platooning of trains along a rail corridor. * We model the problem as a MIP and prove computational complexity. * We evaluate train platooning and compare it with the status quo on real-world data. * Our computational study identifies potential inefficiencies of rail platooning. Abstract Several technological developments are on the way to substantially impact future railway operations. Distributed traction power enables operating longer and heavier trains. The use of automatic couplings allows to quickly join formerly separately operated trains to longer units. Such a platooning on rails promises a much better utilization of congested rail corridors, but requires some upgrades of classical rail infrastructure as well. Rapid shunting tracks are necessary to form platoons and the railway lines have to be prepared to accommodate longer trains. This paper evaluates the rail platooning concept for German railway operator Deutsche Bahn and one of its most-frequented rail corridors. To do so, we formulate the rail platooning problem, which schedules the traversal of freight trains and their assignment to rail platoons along a rail corridor equipped with rapid-shunting facilities. To solve the resulting optimization problem, we apply a three-layered time-space network approach. Our solution procedure optimizing platoon building is benchmarked with conventional (non-platooned) solutions and our results show that considerably more trains can be channeled through a congested rail corridor. Author Affiliation: (a) Friedrich Schiller University of Jena, Department of Management Science, Carl-Zeiß-Straße 3, Jena 07743, Germany (b) University of Passau, Department of Management Science / Operations and Supply Chain Management, Dr.-Hans-Kapfinger-Str. 12, D-94032 Passau, Germany (c) Friedrich Schiller University of Jena, Department of Operations Management, Carl-Zeiß-Straße 3, Jena 07743, Germany * Corresponding author at: Friedrich Schiller University of JenaDepartment of Management ScienceCarl-Zeiß-Straße 3, 07743 Jena, Germany. Article History: Received 13 August 2019; Accepted 5 February 2021 Byline: Stefan Schwerdfeger [stefan.schwerdfeger@uni-jena.de] (*,a,c), Alena Otto [alena.otto@uni-passau.de] (b), Nils Boysen [nils.boysen@uni-jena.de] (c)

Keywords Investment analysis; P2P Lending; Predictive modelling; Ensemble learning; Credit scoring Highlights * Successful Peer-to-Peer (P2P) lending requires an evaluation of loan profitability. * We investigate whether prediction methods and information matter for investment. * We find linear methods perform surprisingly well on several (but not all) criteria. * Ensemble methods outperformance depends on the training measure used. * Using alternative text-based information does not improve profit scoring outcomes * Higher investment returns can be achieved by using linear profitability prediction. Abstract Successful Peer-to-Peer (P2P) lending requires an evaluation of loan profitability from a large universe of loans. Predictions of loan profitability may be useful to rank potential investments. We investigate whether various types of prediction methods and the types of information contained in loan listing features matter for profitable investment. A range of methods and performance metrics are used to benchmark predictive performance, based on a large dataset of P2P loans issued on Lending Club. Robust linear mixed models are used to investigate performance differences between models, according to whether they assume linearity, whether they build ensembles, and which types of predictors they use. The main findings are that: linear methods perform surprisingly well on several (but not all) criteria; whether ensemble methods perform better than individual methods is measure dependent; the use of alternative text-based information does not improve profit scoring outcomes. We conclude that P2P lenders could potentially increase their investment returns by applying linear methods that directly predict the internal rate of return instead of other dependent variables such as loan default. Author Affiliation: (a) Southampton Business School, University of Southampton, Highfield, Southampton SO17 1BJ, UK (b) Risk Analysis & Data Analytics Division, Central Bank of Ireland, North Wall Quay, Dublin 1, Ireland * Corresponding author at: Southampton Business School, University of Southampton, Highfield, Southampton SO17 1BJ, UK. Article History: Received 18 February 2019; Accepted 30 January 2021 (footnote)1 The views expressed in the paper are those of the authors and do not represent the views of the Central Bank of Ireland or the European Central Bank/Single Supervisory Mechanism. The authors would like to thank the reviewers for their constructive suggestions, Lyn Thomas, Leonard Roseman, Justin Silk, Jonathan Crook, and participants at the 2015 Edinburgh Credit Scoring and Credit Control XIV and the 2017 Computational and Financial Econometrics (CFE) Conference for useful discussions on earlier versions of material contained in this paper. They would also like to thank Matt Stevenson for assistance with transfer learning word embedding methods. Byline: Trevor Fitzpatrick [Trevor.Fitzpatrick@soton.ac.uk] (*,1,a,b), Christophe Mues [C.Mues@soton.ac.uk] (a)

Keywords Routing; Scheduling; Inventory costs; Integrated problem; Two-step approach Highlights * The paper considers an integrated scheduling and routing problem with fixed batches, where the objective is to minimize the sum of the inventory costs, the tardiness costs and the routing costs. * Optimal (or near optimal) routes are determined for each batch and each possible departure date using a MILP model, a branch-and-bound and a heuristic algorithm. * These functions are used to minimize the total cost with a new MILP formulation, a CP model and an efficient neighbourhood search method. * Computational experiments are conducted to show the efficiency of the proposed methods. Abstract This paper considers an integrated routing and scheduling problem where the routing part takes into account routing costs and tardiness penalties and the scheduling part is modelled by a permutation flow shop with inventory costs. We assume that each batch is served by a dedicated vehicle, and that the number of batches and their compositions (the number of jobs and the parameters of those jobs) are known in advance. The problem is to determine the starting times of the jobs on each machine in the flow shop, the departure dates of the batches and their delivery route, such that the total cost (sum of inventory, routing and penalty costs) is minimised. A two-step approach is proposed. In a first step, the optimal delivery routes for each batch and each possible departure date are calculated. This is possible as determining the min cost route for a particular delivery batch and a particular departure date is easy. In a second step, we use the delivery cost function for each batch, depending on the departure date, from the first step, to find a schedule that minimises the total cost. Computational experiments are performed on randomly generated instances. Author Affiliation: (a) Université de Tours, LIFAT EA 6300, CNRS, ROOT ERL CNRS 7002, 64 avenue Jean Portalis, 37200 Tours (b) TBS Business School, 1 Place Alfonse Jourdain, 31000 Toulouse * Corresponding author. Article History: Received 31 July 2020; Accepted 3 February 2021 Byline: Hugo Chevroton [hugo.chevroton@univ-tours.fr] (a), Yannick Kergosien [yannick.kergosien@univ-tours.fr] (a), Lotte Berghman [l.berghman@tbs-education.fr] (*,b), Jean-Charles Billaut [jean-charles.billaut@univ-tours.fr] (a)

Keywords OR in energy; Power system resilience; Decision-support systems; Mixed-integer optimization; Wildfires Highlights * Wildfires present important challenges to power system operators. * Power supply security can be improved under renewable energy and wildfire threats. * Wildfire simulation models are coupled with uncertainty sets for renewable energy. * System operators can adapt the network topology to increase wildfire resilience. Abstract Wildfires can pose a major threat to the secure operation of power networks. Chile, California, and Australia have suffered from recent wildfires that have induced considerable power supply cuts. Further, as power systems move to a significant integration of variable renewable energy sources, successfully managing the impact of wildfires on the power supply can become even more challenging due to the joint uncertainty in wildfire trajectories and the power injections from wind and solar farms. Motivated by this, this paper develops a practical decision-support approach that concatenates a stochastic wildfire simulation method with an attacker-defender model that aims to find a worst-case realization for (i) transmission line and generator contingencies, out of those that can potentially be affected by a given wildfire scenario, and for (ii) wind and solar power trajectories, based on a max-min structure where the inner min problem represents a best adaptive response on generator dispatch actions. Further, this paper proposes an evaluation framework to assess the power supply security of various power system topology configurations, under the assumption of limited transmission switching capabilities, and based on the simulation of several wildfire evolution scenarios. Extensive computational experiments are carried out on two representations of the Chilean power network with up to 278 buses, showing the practical effectiveness of the proposed approach for enhancing wildfire resilience in highly renewable power systems. Author Affiliation: (a) Department of Electrical Engineering, Pontificia Universidad Católica de Chile, Santiago 7820436, Chile (b) Department of Industrial and Systems Engineering, Pontificia Universidad Católica de Chile, Santiago 7820436, Chile (c) UC Energy Research Center, Pontificia Universidad Católica de Chile, Santiago 7820436, Chile (d) Department of Economics and Department of Industrial and Systems Engineering, Lehigh University, Bethlehem, PA 18015, USA (e) Faculty of Engineering and Sciences, Universidad Adolfo Ibañez, Santiago, Chile (f) Institute Complex Engineering Systems (ISCI), República 695, Santiago, Chile * Corresponding author at: Department of Electrical Engineering, Pontificia Universidad Católica de Chile, Santiago 7820436, Chile. Article History: Received 12 September 2019; Accepted 2 February 2021 (footnote)[white star] This work was supported by ANID/FONDEF/IT19I0113, SERC-Chile (ANID/FONDAP/15110019), and the Complex Engineering Systems Institute (ANID/FB0816). Byline: Tomás Tapia (a), Álvaro Lorca [alvarolorca@uc.cl] (*,a,b,c), Daniel Olivares (a,c,e,f), Matías Negrete-Pincetic (a,c), Alberto J. Lamadrid L (d)

Keywords Heuristics; Sustainable Supply Chain management; Multi-objective optimization; Lagrangian relaxation; AUGMECON2 Highlights * Designs an efficient Lagrangian matheuristic for a sustainable supply chain problem. * The multi-objective problem has economic, environmental and societal goals. * Competitive with the state-of-the-art exact multi-objective method. * Significantly faster than the exact method. Abstract Sustainable Supply Chain (SSC) management aims at integrating economic, environmental and social goals to assist in the long-term planning of a company and its supply chains. There is no consensus in the literature as to whether social and environmental responsibilities are profit-compatible. However, the conflicting nature of these goals is explicit when considering specific assessment measures and, in this scenario, multi-objective optimization is a way to represent problems that simultaneously optimize the goals. This paper proposes a Lagrangian matheuristic method, called AugMathLagr, to solve a hard and relevant multi-objective problem found in the literature. AugMathLagr was extensively tested using artificial instances defined by a generator presented in this paper. The results show a competitive performance of AugMathLagr when compared with an exact multi-objective method limited by time and a matheuristic recently proposed in the literature and adapted here to address the studied problem. In addition, computational results on a case study are presented and analyzed, and demonstrate the outstanding performance of AugMathLagr. Author Affiliation: (a) Instituto de Ciência e Tecnologia, Universidade Federal de São Paulo, São José dos Campos, Brazil (b) Centro de Estudos de Gestão, Instituto Superior Técnico, Universidade Técnica de Lisboa, Lisboa 1049-101, Portugal * Corresponding author. Article History: Received 13 June 2019; Accepted 4 January 2021 Byline: Camila P.S. Tautenhain [santos.camila@unifesp.br] (a), Ana Paula Barbosa-Povoa [apovoa@tecnico.ulisboa.pt] (b), Bruna Mota [bruna.mota@tecnico.ulisboa.pt] (b), Mariá C.V. Nascimento [mcv.nascimento@unifesp.br] (*,a)

Keywords Forecasting; Forecast combinations; Unbiasedness; Top-down; Hierarchies Highlights * We relate recent literature on Forecast Reconciliation to the extensive body of work on Forecast Combination. * We demonstrate how the linear constraints which naturally apply to the data can be used to generate indirect forecasts of each time-series. These are then combined with direct forecasts to improve forecast accuracy. * The techniques described are generally applicable beyond the hierarchical setting and can improve forecast accuracy in any multivariate forecasting scenario where time-series are subject to linear constraints. * We demonstrate significant improvements in forecast accuracy in the noisiest and hardest to forecast time-series. Abstract A series of recent papers introduce the concept of Forecast Reconciliation, a process by which independently generated forecasts of a collection of linearly related time series are reconciled via the introduction of accounting aggregations that naturally apply to the data. Aside from its clear presentational and operational virtues, the reconciliation approach generally improves the accuracy of the combined forecasts. In this paper, we examine the mechanisms by which this improvement is generated by re-formulating the reconciliation problem as a combination of direct forecasts of each time series with additional indirect forecasts derived from the linear constraints. Our work establishes a direct link between the nascent Forecast Reconciliation literature and the extensive work on Forecast Combination. In the original hierarchical setting, our approach clarifies for the first time how unbiased forecasts for the entire collection can be generated from base forecasts made at any level of the hierarchy, and we illustrate more generally how simple robust combined forecasts can be generated in any multivariate setting subject to linear constraints. In an empirical example, we show that simple combinations of such forecasts generate significant improvements in forecast accuracy where it matters most: where noise levels are highest and the forecasting task is at its most challenging. Author Affiliation: (a) School of Management, University of Bath, United Kingdom (b) Institute for Mathematical Innovation, University of Bath, United Kingdom * Corresponding author. Article History: Received 31 July 2020; Accepted 8 January 2021 Byline: Ross Hollyman [rah98@bath.ac.uk] (*,a), Fotios Petropoulos [f.petropoulos@bath.ac.uk] (a), Michael E. Tipping [mt821@bath.ac.uk] (b)

Keywords Project scheduling; Nash equilibria; Flow networks; Milestones; Lazy-constraint generation Highlights * The paper studies presence of milestones in a multi-agent project scheduling. * We analyze the problem complexity. * The suggested algorithm is based on a lazy-constraint generation technique. * We investigate the role of milestones to drive the project to the owner's goal. Abstract This paper addresses a project scheduling environment in which the activities are partitioned among a set of agents. The project owner is interested in completing the project as soon as possible. Therefore, she/he defines rewards and penalties to stimulate the agents to complete the project faster. The project owner offers a per-day reward for early project completion and defines intermediate project milestones to be met within specific due dates, with associated per-day penalties. Each agent can, therefore, decide the duration of her/his activities, taking into account linear activity crashing costs, the reward for early project completion, and the penalty arising from violating milestone due-dates. We consider Nash equilibria, i.e., situations in which no agent has an interest in individually changing the duration of any of her/his activities, and in particular, the problem of finding a minimum-makespan equilibrium. This problem is known to be NP-hard, and in this paper, we (i) propose a new and efficient exact algorithmic approach for finding the minimum-makespan equilibrium and (ii) through an extensive computational campaign we evaluate the role played by milestones in driving the project towards the owner's goal. Author Affiliation: (a) Czech Institute of Informatics, Robotics, and Cybernetics, Czech Technical University in Prague, Jugoslávských partyzánu 1580/3, Prague 160 00, Czechia (b) Dipartimento di Ingegneria dell'Informazione e Scienze Matematiche, Università degli Studi di Siena, via Roma 56, Siena 53100, Italy (c) LAAS-CNRS, Université de Toulouse, CNRS, UPS, 7 Avenue du Colonel Roche, Toulouse 31400, France * Corresponding author at: Czech Institute of Informatics, Robotics, and Cybernetics, Czech Technical University in Prague, Jugoslávských partyzánu 1580/3, Prague 160 00, Czechia. Article History: Received 8 January 2020; Accepted 13 January 2021 Byline: Premysl Sucha [suchap@cvut.cz] (*,a,c), Alessandro Agnetis [agnetis@diism.unisi.it] (b), Marko Sidlovský (a), Cyril Briand [briand@laas.fr] (c)

Keywords Data envelopment analysis; Exponential distance function; Technical efficiency; Log-convexity; Cobb-Douglas structure Highlights * Production theoretic basis for the exponential distance function is presented. * Napierian technology and the corresponding profit function are defined. * Duality between the exponential distance and the Napierian profit functions. * Shadow prices based on Cobb-Douglas structure are derived. * A notion of log-constant returns to scale is introduced. Abstract Charnes, Cooper, Seiford, and Stutz (1982, 1983) and Banker and Maindiratta (1986) suggested multiplicative radial measures for efficiency gauging. More recently, Peyrache and Coelli (2009) presented a multiplicative directional distance function that was elaborated by Mehdiloozad, Sahoo, and Roshdi (2014). In this paper, we extend these studies and propose an exponential approach based upon a new exponential distance function endowed with a multiplicative production structure. The main purposes of this paper are twofold: one is to provide a general production theoretic basis for the approach and the other is to extend it to a nonparametric framework. The first purpose is accomplished as follows: (1) the exponential distance function is formally defined and its properties are established; (2) it is shown how the exponential distance function is characterized under a Napierian technology; (3) a duality relationship between Napierian profit and the exponential distance functions is established; (4) shadow prices of inputs and outputs are derived based on the Napierian technology. The second purpose is accomplished by providing nonparametric programming extensions, which include data envelopment analysis (DEA) models, productivity indexes and returns to scale models. Here, the efficacy of our nonparametric theoretical results is demonstrated by applying DEA to the data on accommodation establishments in OECD. Author Affiliation: (a) University of Perpignan, Department of Economics, LAMPS, 52 avenue Paul Alduy, Perpignan 66000, France (b) Faculty of Commerce, Fukuoka University, Fukuoka City 814-0143, Japan (c) University of Perpignan, Department of Economics, MRM, 52 avenue Paul Alduy, Perpignan 66000, France * Corresponding author. Article History: Received 5 November 2019; Revised 22 July 2020; Accepted 24 November 2020 (footnote)1 The authors would like to thank the editor and three anonymous reviewers for their valuable comments and suggestions that helped to substantially improve the paper. The second author was partially supported by funding from Fukuoka University (Grant No. 204007). Byline: Walter Briec (a), Hirofumi Fukuyama [fukuyama@fukuoka-u.ac.jp] (*,b), Paola Ravelojaona (1,c)

Keywords Finance; Investment under uncertainty; Time to build; Market incompleteness; Risk aversion Highlights * Develop theoretical framework of investment under uncertainty with a time to build. * Market incompleteness and risk aversion are incorporated into the framework. * Results on market incompleteness and risk aversion challenged by time to build. * Time to build models ought to incorporate market incompleteness and risk aversion. Abstract In this paper I develop a theoretical framework of irreversible investment under uncertainty in which there is time to build (TTB). The novel aspects of this framework, compared with TTB models in the extant literature, are that the market is incomplete in that not all the uncertainty associated with investing can be diversified away by trading an appropriate spanning asset, and the decision-maker, who acts in the interest of an organisation, is risk averse. I show that models of investment under uncertainty with a TTB, and models of investment under uncertainty with market incompleteness and risk aversion, ought not to be mutually exclusive as they have been in research to date because the recognised results of market incompleteness and risk aversion on the optimal investment strategy are challenged when we incorporate a TTB feature. Conversely, there are also implications on the effect of a TTB when we incorporate market incompleteness and risk aversion. The framework I develop in this paper provides a robust and parsimonious means of facilitating this. Author Affiliation: Department of Economics, City University, Northampton Square, London EC1V 0HB, UK Article History: Received 5 August 2020; Accepted 26 December 2020 (footnote)1 ORCID ID: https://orcid.org/0000-0003-0944-9894 Byline: Laura Delaney [laura.delaney.1@city.ac.uk] (1)

Keywords Excess kurtosis; Skewness; Stochastic frontier model; Specification testing Highlights * Empirical skewness and kurtosis can serve as useful diagnostics. * Present maximal bounds for skewness/kurtosis of error from stochastic frontier model. * Construct test of correct distributional assumptions. * Provide several empirical applications of test. Abstract The distributional specifications for the composite regression error term most often used in stochastic frontier analysis are inherently bounded as regards their skewness and excess kurtosis coefficients. We derive general expressions for the skewness and excess kurtosis of the composed error term in the stochastic frontier model based on the ratio of standard deviations of the two separate error components as well as theoretical ranges for the most popular empirical specifications. While these simple expressions can be used directly to assess the credibility of an assumed distributional pair, they are likely to over reject. Therefore, we develop a formal test based on the implied ratio of standard deviations for the skewness and the kurtosis. This test is shown to have impressive power compared with other tests of the specification of the composed error term. We deploy this test on a range of well-known datasets that have been used across the efficiency community. For many of them we find that the classic distribution assumptions cannot be rejected. Author Affiliation: (a) Department of Economics, Athens University of Economics and Business, 28 Oktovriou 76, Athina 104 34, Greece (b) Department of Economics, University of Miami, Coral Gables, FL 33146, USA * Corresponding author. Article History: Received 4 April 2020; Revised 29 December 2020; Accepted 31 December 2020 (footnote)[white star] This paper was previously circulated under the titles 'Some Skewed Results on the Stochastic Frontier Model' and 'Moment Diagnostics and Quasi-Maximum Likelihood Estimation for the Stochastic Frontier Model.' We thank participants at EWEPA 2019 for valuable feedback as well as David Bernstein, William Horrace, Subal Kumbhakar, Peter Schmidt and Ian Wright for useful comments which improved the paper. The comments and criticisms of the editor and three anonymous reviewers are greatly appreciated as well. The usual disclaimer applies. All R and gretl code used in this paper is available upon request. Byline: Alecos Papadopoulos [papadopalex@aueb.gr], Christopher F. Parmeter [cparmeter@bus.miami.edu] (*)

Keywords Decision analysis; Investment analysis; Capital investment; Smooth ambiguity model; Technology shock Highlights * This paper analyzes the behavior of a firm facing an ambiguous technology shock. * Under some conditions, we show that the optimal amount of capital investment increases (decrease) in ambiguity aversion if the production function is substitute (complement). * We show that this result is counterintuitive when the production function is substitute. Abstract This paper analyzes the behavior of a firm facing an ambiguous technology shock and the effects of the attitude toward ambiguity on optimal capital investment using the smooth ambiguity model of Klibanoff et al. (2005). Although it seems intuitive that an increase in ambiguity aversion always reduces the optimal capital investment, this is not necessarily true because the shape of the production function plays a key role in determining the effect. Under some conditions, we show that the optimal amount of capital investment increases (decreases) in ambiguity aversion if the production function is substitute (complement), and that this result is counterintuitive when the production function is substitute. Furthermore, our main results hold if we assume the [alpha]-maxmin preferences in Ghirardato et al. (2004). Author Affiliation: (a) Faculty of Economics, Okayama University, 3-1-1 Tsushimanaka, Kita-ku, Okayama 700-8530, Japan (b) Faculty of Commerce, Waseda University, 1-6-1 Nishiwaseda, Shinjyku-ku, Tokyo 169-8050, Japan * Corresponding author. Article History: Received 15 January 2020; Revised 5 October 2020; Accepted 27 November 2020 (footnote)[white star] We acknowledge the Editor (Emanuele Borgonovo) and an anonymous reviewer, whose comments improved this paper substantially. We are grateful to Shin Kanaya, Kazuko Kano, Hajime Katayama, Hideyuki Mizobuchi, Michio Suzuki, Akihisa Shibata, Kit Pong Wong, and participants at the second Kwansei Gakuin University and KIER joint workshop for their comments and discussions. This research is financially supported by the JSPS KAKENHI Grant Numbers 20K01745, 17K03806, 20H01507, and 26705004, and the Joint Research Program of KIER. Byline: Takao Asano [asano@e.okayama-u.ac.jp] (*,a), Yusuke Osaki [osakiy@waseda.jp] (b)

Keywords OR in maritime industry; Tandem quay crane; Container slot planning; Truck scheduling; Joint optimization Highlights * A new integrated optimization problem of container terminal to fully utilize the tandem quay crane (TQC) is proposed. * Both container slot planning and truck scheduling are illustrated to have significant impacts on the efficiency of the TQC. * A novel model for integrated optimization of container slot planning and truck scheduling based on the TQCs is developed. * A greedy randomized adaptive search procedure is designed to solve the problem. * The effectiveness of the proposed methods and the significance of the integrated optimization are verified. Abstract This paper examines the loading operations of a new type of quay crane, the tandem quay crane (TQC), which has been designed to increase the terminal productivity by lifting more containers simultaneously. Due to the special characteristics of the TQC, more than one truck is required to serve a tandem-lift of it, and the containers handled in a tandem-lift must be loaded into two neighboring rows in the same tier of the containership. Of particular importance in schedules at the terminal is to coordinate the truck scheduling and container stowage slot plan with the TQCs, which is a key challenge in practice. This paper is an attempt to tackle this issue by developing a mixed integer linear programming (MILP) model with an objective of minimizing the completion time of the loading operation under the full utilization of the TQC. Due to the complexity of the MILP model, we then propose a model to derive lower bounds of the problem, and a greedy randomized adaptive search procedure (GRASP) to solve the problem. Computational experiments are conducted with a variety of instances. The results derived by GRASP for solving small size problems are within 0.04% of the optimal results obtained by the Gurobi. For large scale instances, GRASP outperforms the Gurobi in terms of the solution quality and computation time. Additionally, on average results derived by GRASP are within 8.2% of the lower bound. Further experiments demonstrate the advantage of the integrated optimization of container slot planning and truck scheduling for TQCs. Author Affiliation: Department of Transportation Engineering, Dalian Maritime University, Linghai Road No.1th, Ganjingzi District, Dalian 116026, Liaoning, China * Corresponding author. Article History: Received 10 December 2019; Revised 29 November 2020; Accepted 2 December 2020 Byline: Lingrui Kong, Mingjun Ji [jmj@dlmu.edu.cn] (*), Zhendi Gao [356190646@qq.com]

Keywords Finance; Credit risk modelling; Discrete Markov chains; Lumpability Highlights * The paper studies the problem of compliance of internal credit rating models used for pricing and provisioning. * We apply the approximate lumpability method to reduce a Markov chain to an equivalent, lower-dimensional process. * An algorithm to solve the problem of approximate lumpability is proposed, which, often reduces to a closed form solution. * The proposed methodology is applied to multiple examples of credit risk modelling and applications. * The proposed techniques are applicable to other problems related to Markov chain modelling e.g. decision theory and finance. Abstract The aim of this paper is threefold. Firstly, we define the necessary quantities associated to the lumpability of a Markov chain and study their fundamental properties. Secondly, we examine the case of approximate lumpability of a non-lumpable Markov and an efficient method of minimizing the error in the approximation. Finally, we introduce a family of general minimization problems that can be approached using this method and examine applications in credit risk modelling, particularly under recent regulatory changes related to loan classification and provision calculations under IFRS 9. Author Affiliation: (a) Department of Statistics and Stochastic Modelling and Applications Laboratory, Athens University of Economics and Business, 76 Patission Str, 10434 Athens, Greece (b) Department of Mathematics, University of Sussex, Pevensey 2, BN1 9QH Brighton, UK * Corresponding author. Article History: Received 21 October 2019; Revised 8 November 2020; Accepted 10 November 2020 Byline: K. Georgiou (a), G.N. Domazakis (b), D. Pappas (a), A.N. Yannacopoulos [ayannaco@aueb.gr] (*,a)