Search

The booming development of customised e-commerce makes e-commerce supply chains experience a severe test of consumer service level (CSL). An efficient e-commerce supply chain resilience optimisation method addressing economic performance and service performance is proposed to improve customised services under disruption risks. The proposed method includes a hybrid strategy considering both a resistance strategy with blockchain (BCT) adoption and time-dependent recovery strategies simultaneously. A two-stage multi-period multi-product stochastic programming with BCT adoption is then proposed, which (1) implements BCT as the resistance strategy in the pre-disruption stage; (2) collaborates four recovery strategies in the post-disruption stage; (3) supports the e-tailer in making decisions and optimises his profit and order fulfillment time while taking product priorities into account. Using the actual data of Chinese e-commerce during the public health emergency in 2020, it is demonstrated that (1) the applicability of the model with BCT adoption in both supply chain resilience and CSL improvement; (2) the performance of the hybrid strategy in long-term disruption management; (3) the efficiency and robustness of the proposed solution approach for multi-objective high-dimensional stochastic programming problems. E-tailers can optimise decision-making under long-term disruptions by the proposed methodology in practices. [ABSTRACT FROM AUTHOR]

This paper delves into the impacts of an ongoing global crisis on the resilience of supply chains. Furthermore, it proposes measures to address and mitigate the disruptions caused by the prevailing uncertainties. For example, while the economy has started to recover after the pandemic and demand has increased, companies have not fully returned to their pre-pandemic levels. To enhance their supply chain resilience and effectively manage disruptions, one viable strategy is the implementation of flexible/hybrid manufacturing systems. This research is motivated by the specific requirements of Vestel Electronics, a household appliances company, which seeks a flexible/hybrid manufacturing production setup involving dedicated machinery to meet regular demand and the utilisation of flexible manufacturing system (FMS) to handle surges in demand. We employ a scenario-based approach to model demand uncertainty, enabling the company to make immediate and adaptive decisions that take advantage of the cost-effectiveness of standard production and the responsiveness of FMS. To solve the problem, we propose a heuristic algorithm based on column generation. The numerical results demonstrate that our optimisation model provides solutions with an average optimality gap of less than 6% while also reducing the average cost of standard production schemes without FMS by over 12%. [ABSTRACT FROM AUTHOR]

Two of the major problems Traditional Manufacturing (TM) supply chains face are setting requisite reactive strategies to address the uncertainties in demand and the optimal placement of these buffering capacities in order to be both responsive and cost-effective. With Additive Manufacturing (AM) stepping into large-scale production at different firms, we address the aforementioned supply chain dilemmas by considering the potential role of AM in a TM supply chain network where AM facilities can act as a recourse to the TMs and, also, as a dedicated source providing responsive and cost-effective sourcing alternatives. We develop an analytical allocation rule based on cost differentials, which provides optimal sourcing decisions through sequential demand replenishment and facilitates an efficient performance evaluation of possible network configurations. We first model the scenario as a three-stage stochastic optimisation problem. We then solve it using the allocation rule and present an illustration of our analysis and the optimal supply chain network configuration. Furthermore, we derive some insights as to how different problem characteristics affect the value and usage of AM. [ABSTRACT FROM AUTHOR]

This paper addresses the limited adaptability and the computational burden of energy management systems (EMSs) for hybrid electric vehicles (HEVs) implemented via dynamic programming (DP)-based approaches. First, a deterministic dynamic programming (DDP) framework is presented to solve HEV EMS problems subject to a specific driving cycle. To address this limitation, an improved DDP approach, integrating the actual travelled position of the vehicle into the control law, is proposed. This way, a given DDP-based EMS can be applied to all the driving cycles, yet still measured on the same road. Stochastic dynamic programming (SDP)-based EMSs are also developed and prove to be more adaptive to driving scenarios completely different from the ones used for their computation. Real-world driving cycles are employed in all the presented cases, while a reduced HEV powertrain model is used to alleviate the typical DP computational burden. [ABSTRACT FROM AUTHOR]

Optimal design facilitates intelligent data collection. In this paper, we introduce a fully Bayesian design approach for spatial processes with complex covariance structures, like those typically exhibited in natural ecosystems. Coordinate exchange algorithms are commonly used to find optimal design points. However, collecting data at specific points is often infeasible in practice. Currently, there is no provision to allow for flexibility in the choice of design. Accordingly, we also propose an approach to find Bayesian sampling windows, rather than points, via Gaussian process emulation to identify regions of high design efficiency across a multi-dimensional space. These developments are motivated by two ecological case studies: monitoring water temperature in a river network system in the northwestern United States and monitoring submerged coral reefs off the north-west coast of Australia. [ABSTRACT FROM AUTHOR]

An open-source modelling framework Predicer, standing for Predictive Decider, for multi-market day-ahead market operation purposes is described in this paper. The Predicer model uses scenario-based stochastic optimisation to obtain decision variables and bid matrixes for energy and reserve markets by maximising the risk-adjusted expected value of the profit during the model time frame. The modelled energy system structure is abstract, that is, based on basic elements such as nodes representing different energy types and processes representing flows between nodes. The abstract model structure enables user to construct arbitrary energy systems and define links between assets, commodities, energy markets and reserve markets. Predicer model can include properties such as unit ramp rates, online units, dynamic energy storages, market realisation and market bidding requirements. The aggregation of unit-based energy and reserve opportunities into a virtual power plant allows the asset owner to make optimized portfolio-level bids for different market products. The model scenarios consist of user defined forecasts for market prices, renewable energy supply, energy demand and other system related time series. Predicer is implemented in Julia programming language and uses the JuMP optimisation package. [ABSTRACT FROM AUTHOR]

The buffer allocation problem in production lines is an NP-hard combinatorial optimisation problem. This paper proposes a new hybrid optimisation approach (using simulation) relying on genetic algorithm (GA) and finite perturbation analysis (FPA). Unlike the infinitesimal perturbation analysis, which deals with small (infinitesimal variation) perturbations for estimating gradients of the performance measure, FPA deals with larger (finite) or more lasting perturbations. It is an extension specifically dedicated to discrete decision variables and applicable to most discrete-event dynamic systems. The proposed method allows a global search using GA, with refinement in specific solution-space regions using FPA. The main objective is to maximise the average production rate of a production line with unreliable machines, by allocating the total buffer capacity in locations between machines. Extensive numerical experiments show that: (1) the proposed hybrid GA-FPA method clearly outperforms the state-of-the-art methods from the literature; (2) combining FPA and GA is beneficial when compared to employing GA or FPA independently. [ABSTRACT FROM AUTHOR]

We introduce and investigate stochastic processes designed to find local minimizers and saddle points of non-convex functions, exploring the landscape more efficiently than the standard noisy gradient descent. The processes switch between two behaviours, a noisy gradient descent and a noisy saddle point search. It is proven to be well-defined and to converge to a stationary distribution in the long time. Numerical experiments are provided on low-dimensional toy models and for Lennard–Jones clusters. [ABSTRACT FROM AUTHOR]

Stochastic scheduling has received much attention from both industry and academia. Existing works usually focus on random job processing times. However, the uncertainty existing in job release times may largely impact the performance as well. This work investigates a stochastic parallel machine scheduling problem, where job release times and processing times are uncertain. The problem consists of a two-stage decision-making process: (i) assigning jobs to machines on the first stage before the realisation of uncertain parameters (job release times and processing times) and (ii) scheduling jobs on the second stage given the job-to-machine assignment and the realisation of uncertain parameters. The objective is to minimise the total cost, including the setup cost on machines (induced by job-to-machine assignment) and the expected penalty cost of jobs' earliness and tardiness. A two-stage stochastic program is proposed, and the sample average approximation (SAA) method is applied. A scenario-reduction-based decomposition approach is further developed to improve the computational efficiency. Numerical results show that the scenario-reduction-based decomposition approach performs better than the SAA, in terms of solution quality and computation time. [ABSTRACT FROM AUTHOR]

The simple assembly line balancing (SALB) problem is a significant challenge faced by industries across various sectors aiming to optimise production line efficiency and resource allocation. One important issue when the decision-maker balances a line is how to keep the cycle time under a given time across all cells, even though there is variability in some parameters. When there are stochastic elements, some approaches use constraint relaxation, intervals for the stochastic parameters, and fuzzy numbers. In this paper, a three-part algorithm is proposed that first solves the balancing problem without considering stochastic parameters; then, using simulation, it measures the effect of some parameters (in this case, the inter-arrival time, processing times, speed of the material handling system which is manually performed by the workers in the cell, and the number of workers who perform the tasks on the machines); finally, the add-on OptQuest in SIMIO solves an optimisation problem to constrain the cycle time using the stochastic parameters as decision variables. A Gearbox instance from literature is solved with 15 tasks and 14 precedence rules to test the proposed approach. The deterministic balancing problem is solved optimally using the open solver GLPK and the Pyomo programming language, and, with simulation, the proposed algorithm keeps the cycle time less than or equal to 70 s in the presence of variability and deterministic inter-arrival time. Meanwhile, with stochastic inter-arrival time, the maximum cell cycle is 72.04 s. The reader can download the source code and the simulation models from the GitHub page of the authors.

The simple assembly line balancing (SALB) problem is a significant challenge faced by industries across various sectors aiming to optimise production line efficiency and resource allocation. One important issue when the decision-maker balances a line is how to keep the cycle time under a given time across all cells, even though there is variability in some parameters. When there are stochastic elements, some approaches use constraint relaxation, intervals for the stochastic parameters, and fuzzy numbers. In this paper, a three-part algorithm is proposed that first solves the balancing problem without considering stochastic parameters; then, using simulation, it measures the effect of some parameters (in this case, the inter-arrival time, processing times, speed of the material handling system which is manually performed by the workers in the cell, and the number of workers who perform the tasks on the machines); finally, the add-on OptQuest in SIMIO solves an optimisation problem to constrain the cycle time using the stochastic parameters as decision variables. A Gearbox instance from literature is solved with 15 tasks and 14 precedence rules to test the proposed approach. The deterministic balancing problem is solved optimally using the open solver GLPK and the Pyomo programming language, and, with simulation, the proposed algorithm keeps the cycle time less than or equal to 70 s in the presence of variability and deterministic inter-arrival time. Meanwhile, with stochastic inter-arrival time, the maximum cell cycle is 72.04 s. The reader can download the source code and the simulation models from the GitHub page of the authors. [ABSTRACT FROM AUTHOR]

Due to gradual disappearance of global demographic dividend, the improvement of workforce efficiency becomes increasingly important. It has been commonly recognised that workforce motivation can largely stimulate the workers, especially in manufacturing systems. This paper investigates a worker scheduling problem under given work shifts with workforce motivation effects, where motivation effects are characterised by motivation coefficients of job processing times. We focus on the situation where motivation coefficients are uncertain due to various factors, and only the mean vector and covariance matrix are known. The objective is to maximise the service level, measured by the probability of ensuring no tardy jobs. We first propose a distributionally robust chance constrained formulation with a probabilistic objective function. Then an adapted sample average approximation (SAA) method and a heuristic, based on an approximated mixed integer second-order cone programming (MI-SOCP) model and the idea of problem decomposition, is developed. Numerical results show that the decomposition-based heuristic is more efficient. We also draw some managerial insights. [ABSTRACT FROM AUTHOR]

Disassembly line balancing problem (DLBP), which is to select disassembly process, open workstations and assign selected tasks to opened workstations, plays an important role in the recycling of End Of Life products. In real-world disassembly operations, task processing times are usually stochastic due to various factors. Most related works address the uncertain processing times by assuming that the probability distribution is known and the task processing times are independent of each other. In practice, however, it is difficult to get the complete distributional information and there is always underlying correlation between the uncertain processing times. This paper investigates the DLBP with partial uncertain knowledge, i.e. the mean and covariance matrix of task processing times. A new distributionally robust formulation with a joint chance constraint is proposed. To solve the problem, an approximated mixed integer second-order cone programming (MI-SOCP) model is proposed, and a two-stage parameter-adjusting heuristic is further developed. Numerical experiments are conducted, to evaluate the performance of the proposed method. We also draw some managerial insights and consider an extension problem. [ABSTRACT FROM AUTHOR]

Phase stability and equilibrium calculations are very significant in the process industries, ranging from process simulation, equipment design, operation and design. These problems involve several systems of complex and nonlinear equations, which are mathematically challenging. An alternative solution is formulating the problems into a global optimisation problem. A robust and reliable optimisation technique is required to solve these global optimisation problems. In this study, two recently developed stochastic global optimisation algorithms, POA and GTO were applied to phase stability and equilibrium problems. The need to enhance their performance necessitated the development of two modified algorithms, PGOA1 and PGOA2, which are the hybridisation of POA and GTO. All the algorithms' abilities were explored for solving computationally challenging and multidimensional phase stability and equilibrium problems, and their performances were compared. The implementation of POA and GTO was good in solving most problems, though GTO slightly outperformed POA. The hybridisation of both to develop PGOA1 and PGOA2 was worthwhile as both outperformed the POA and GTO. PGOA2 is more robust and reliable in solving phase stability and equilibrium, and its performance superseded the performances of the other algorithms. • Pelican and Gorilla troops algorithms applied to phase problems. • Phase stability and equilibrium for non-reactive and reactive benchmark problems. • Hybrid of Pelican and Gorilla troop algorithms outperformed the parent algorithms. • The inclusion of a local optimizer improved the algorithms. [ABSTRACT FROM AUTHOR]

Stochastic scheduling optimisation is a hot and challenging research topic with wide applications. Most existing works on stochastic parallel machine scheduling address uncertain processing time, and assume that its probability distribution is known or can be correctly estimated. This paper investigates a stochastic parallel machine scheduling problem, and assumes that only the mean and covariance matrix of the processing times are known, due to the lack of historical data. The objective is to maximise the service level, which measures the probability of all jobs jointly completed before or at their due dates. For the problem, a new distributionally robust formulation is proposed, and two model-based approaches are developed: (1) a sample average approximation method is adapted, (2) a hierarchical approach based on mixed integer second-order cone programming (MI-SOCP) formulation is designed. To evaluate and compare the performance of the two approaches, randomly generated instances are tested. Computational results show that our proposed MI-SOCP-based hierarchical approach can obtain higher solution quality with less computational effect. [ABSTRACT FROM AUTHOR]

In the context of simulation-based optimisation, this paper reviews recent work related to the role of metaheuristics, matheuristics (combinations of exact optimisation methods with metaheuristics), simheuristics (hybridisation of simulation with metaheuristics), biased-randomised heuristics for 'agile' optimisation via parallel computing, and learnheuristics (combination of statistical/machine learning with metaheuristics) to deal with NP-hard and large-scale optimisation problems in areas such as transport and logistics, manufacturing and production, smart cities, telecommunication networks, finance and insurance, sustainable energy consumption, health care, military and defence, e-marketing, or bioinformatics. The manuscript provides the main related concepts and updated references that illustrate the applications of these hybrid optimisation–simulation–learning approaches in solving rich and real-life challenges under dynamic and uncertainty scenarios. A numerical analysis is also included to illustrate the benefits that these approaches can offer across different application fields. Finally, this work concludes by highlighting open research lines on the combination of these methodologies to extend the concept of simulation-based optimisation. [ABSTRACT FROM AUTHOR]

Copyright of LHB: Hydroscience Journal is the property of Taylor & Francis Ltd and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

We propose a modification of the pure random search algorithm for cases when the global optimum point can be located near the boundary of a feasible region. If the feasible region is cube-shaped, the worst case occurs when the global optimum point is located at the vertex of a cube. In these cases, the sample size for the pure random search should be increased significantly because the success probability of the single trial is decreased. The proposed modification is based on the virtual extension of the feasible region. The extended region contains ε -neighbourhoods of all points of the original region. All random points that fall outside the original feasible region are mapped to the nearest points on its boundary before the values of an objective function at these points are calculated. This extension of the feasible region also leads to an increase in the sample size due to the increased volume. We compare sample sizes required to hit into the neighbourhood of the global optimum point with a given probability for the proposed method and the original pure random search algorithm with a corrected sample size. We consider ball-shaped and cube-shaped feasible regions and find the sufficient conditions for the proposed algorithm to be more efficient. We offer a practical recommendation for cases, in which the feasible region is defined by a system of linear inequalities. [ABSTRACT FROM AUTHOR]

Grinding mills exhibit steady-state behaviour, formalised by quadratic performance functions called grind curves. Grind curves are predominantly affected by feed-ore characteristics that are generally unknown during regular operation. This work assumes that optimising grinding mills, using grind curves, is akin to bandit problems that can be solved using continuum-armed bandit algorithms. This paper proposes a constrained continuum-armed bandit agent based on Thompson sampling that uses linear models to model the agent reward as a combination of real-value and binary components. Real-value reward components represent mill-performance indicators and are modelled as second-order polynomials of the agent's actions, thereby approximating the unknown grind curves. Binary components indicate constraint violations in the environment and are used to train linear constraint models for the optimisation algorithm selecting the optimal actions. Agent testing occurred in an environment containing a dynamic grinding mill model calibrated to grind curve data. Optimisation performance was compared to extremum-seeking controllers operating in the same environment. The experiments demonstrate that the continuum-armed bandit agent could find the unknown grind curves and steer the process near the optimal operating point. The continuum-armed bandit agent followed similar trajectories to the extremum-seeking controllers in most experiments. The agent requires tuning a smaller number of hyper-parameters than the extremum-seeking controllers, with physically meaningful interpretations. The agent improved upon the optimisation performance of one variant of the extremum-seeking controllers by 2% while producing a 1% decrease compared to the remaining two controllers. The agent proved robust to algorithm hyper-parameters, obtaining similar rewards for various agent configurations. [ABSTRACT FROM AUTHOR]

This paper studies the impact of distributed energy resources (DER) on economic displacements or delays of power system investments. We investigate how the operational flexibility from DER controllability can influence the integrated expansion of transmission and energy storage. Given the interplay of flexibility provision from different technologies, accurately representing uncertainties is essential to avoid over- or under-estimating the flexible operational capabilities of DER. To address this challenge, we propose a multi-stage stochastic expansion planning model that can optimise transmission and storage investments, as well as DER services against long-term uncertainties and detailed operational constraints. We employ a four-stage scenario tree to represent uncertainties and a Dantzig–Wolfe decomposition within a column generation approach to tackle computational challenges. Case studies performed on real Australian National Electricity Market (NEM) scenarios demonstrate that a deterministic model overestimates the capabilities of controllable DER to displace transmission investments, particularly in early stages. Conversely, the proposed stochastic model provides a more measured assessment, maintaining a steadier estimate of transmission displacement potential by controllable DER throughout various stages. • Multi-stage model optimising DER with investments in transmission and storage. • Analysis of DER controllability's impact on investments under uncertainty. • Comparison of deterministic and stochastic approaches in real scenarios. • Stochastic model enhances investment robustness by incorporating flexible DER. • Deterministic models show increased uncertainty in resulting portfolios. [ABSTRACT FROM AUTHOR]

Security-constrained transmission expansion planning (SCTEP) is an inherently complex problem that requires simultaneously solving multiple contingency states of the system (usually corresponding to N-1 security criterion). Existing studies focus on effectively finding optimal solutions; however, single optimal solutions are not sufficient to interpret the value of flexibility (e.g. , from energy storage systems) and support system planners in well-informed decision making. In view of planning uncertainties, it is necessary to estimate the contributions of flexibility to various objectives and prioritise the most effective investments. In this regard, this work introduces a SCTEP tool that enables interpreting the value of flexibility in terms of contributions to avoided load curtailment and total expected system cost reduction. Inspired by cooperative game theory, the tool ranks the contributions of flexibility providers and compares them against traditional line reinforcements. This information can be used by system planners to prioritise investments with higher contributions and synergistic capabilities. [ABSTRACT FROM AUTHOR]

This paper investigates the impact of implementing single and multi-optimisation solutions on the biological treatment process in a sequencing batch reactor (SBR). The research is based on a case study of the water resource recovery facility (WRRF) in Swarzewo, Northern Poland. The paper introduces the adaptive extremum seeking control (ESC) method for dissolved oxygen (DO) concentration control and places it in a layered control structure. Further, it presents the introduction of an optimisation layer for the structure and parameters of the SBR cycle, through the synthesis of stochastic methods: single-objective optimisation (SOO) using a genetic algorithm (GA) and multi-objective optimisation (MOO) using the NSGA-II algorithm. The results were compared to a classical approach with fixed cycle parameters. The paper shows the advantages of optimising cycle parameters, including the number of phases as well as the DO value, on the process flow. These control structures underwent simulation tests in the MATLAB environment with the Simba package. The biochemical processes occurring in the reactor are based on the Activated Sludge Model No. 2d (ASM2d). The optimising control system demonstrates tangible improvements in operational efficiency and significant reductions in electrical energy consumption, highlighting the effectiveness of the proposed methodologies. © 2017 Elsevier Inc. All rights reserved. • Wastewater treatment is one of the key concerns in ecology and environmental engineering. • Hierarchical control system has been used to the WWTP. • Optimisation of biological processes reduces WWTP operating costs. [ABSTRACT FROM AUTHOR]

The current energy transition needed to meet the world climate objectives is causing stability challenges in all the power systems. As a consequence, finding solutions that allow dealing with those frequency stability issues is critical to achieving the sustainability objectives. In this context, batteries can find new revenue opportunities by being part of the solution to this problem. This paper develops a mathematical model that provides the optimal bidding strategy and the most convenient operational planning for batteries participating in balancing markets. More specifically, this project is focused on the Swedish frequency markets whose minimum bid size is smaller than 5 MW. The proposed optimization problem is a two-stage stochastic mixed-integer and linear optimization from the battery operator’s point of view that includes a novel step-by-step process for properly selecting the bid prices by the quantification of the risks, control costs and amortization of the battery. The originality of the work resides in the aforementioned process that allows risk assessment together with the linear modelling of paid-as-bid markets, commonly solved through non-linear problems. As a result, not only the profitability of this application for batteries is demonstrated, but it is also possible to observe the seasonal differences when it comes to revenue and power requirements. The model is tested with data from Sweden, but it is designed to be adjustable to other balancing markets.

Ageneric class of multistage optimisation problems related to production/inventorymanagement underMarkovian uncertainty is introduced and investigated. For each instance in the class, it is shown how to construct state-space representable uncertainty sets at any probability level, thus leading to efficient resolution of both the stochastic and robust versions of the problem. Computational experiments aimed at comparing the optimal strategies corresponding to both versions in terms of risk are then reported and discussed; it is observed that the robust optimisation approach can significantly outperform the stochastic optimisation approach when targeting lower risk levels (typically less than 2%). [ABSTRACT FROM AUTHOR]

We present and numerically analyse the Basin Hopping with Skipping (BH-S) algorithm for stochastic optimisation. This algorithm replaces the perturbation step of basin hopping (BH) with a so-called skipping mechanism from rare-event sampling. Empirical results on benchmark optimisation surfaces demonstrate that BH-S can improve performance relative to BH by encouraging non-local exploration, that is, by hopping between distant basins. [ABSTRACT FROM AUTHOR]

Metaheuristics provide the means to approximately solve complex optimisation problems when exact optimisers cannot be utilised. This led to an explosion in the number of novel metaheuristics, most of them metaphor-based, using nature as a source of inspiration. Thus, keeping track of their capabilities and innovative components is an increasingly difficult task. This can be resolved by an exhaustive classification system. Trying to classify metaheuristics is common in research, but no consensus on a classification system and the necessary criteria has been established so far. Furthermore, a proposed classification system can not be deemed complete if inherently different metaheuristics are assigned to the same class by the system. In this paper we provide the basis for a new comprehensive classification system for metaheuristics. We first summarise and discuss previous classification attempts and the utilised criteria. Then we present a multi-level architecture and suitable criteria for the task of classifying metaheuristics. A classification system of this kind can solve three main problems when applied to metaheuristics: organise the huge set of existing metaheuristics, clarify the innovation in novel metaheuristics and identify metaheuristics suitable to solve specific optimisation tasks. [ABSTRACT FROM AUTHOR]

With rising energy prices, it becomes increasingly important for industrial companies to consider energy costs in the activity of production planning and control. Reduced energy costs can both be realised through lower consumption and higher energy efficiency. Furthermore, the ability to flexibly consume energy, while adapting consumption to electricity availability or variable energy prices, as often referred to as 'demand response', might bring significant benefits. Two important requirements for demand response are the ability to predict consumption and the ability to precisely determine the potential flexible approaches available. For this purpose, a material flow model was implemented that simulates both the production process and the energy consumed by it. This model was parameterised based on real production data and energy consumption. Initial analysis reveals that manufacturing processes generally undergo interruptions, which cannot be predicted at the stage of production planning. These interruptions significantly reduce the accuracy of the model and were considered by means of probability density functions. In this paper, the forging shop of an automotive supplier is used as a use case. The main power consumption in this concrete case is caused by induction furnaces, which require different amounts of electricity, depending on the workpiece. Consequently, the loads can be adjusted by adapting the sequence of production orders. Using the simulation model developed, the peak loads of the forging shop were analysed, and an energy-orientated production plan was created and optimised using a genetic algorithm. Results show that the peak loads of the plant can be reduced significantly (by more than 10 %), despite the uncertainties of production, by means of stochastic modelling and optimisation. The application of these findings might contribute to both the stability of the electricity grid and the reliability of operations. [ABSTRACT FROM AUTHOR]

This paper deals with the selective maintenance problem for a multi-component system performing consecutive missions separated by scheduled breaks. To increase the probability of successfully completing its next mission, the system components are maintained during the break.Alist of potential imperfect maintenance actions on each component, ranging from minimal repair to replacement is available. The general hybrid hazard rate approach is used to model the reliability improvement of the system components. Durations of the maintenance actions, the mission and the breaks are stochastic with known probability distributions. The resulting optimisation problem is modelled as a non-linear stochastic programme. Its objective is to determine a cost-optimal subset of maintenance actions to be performed on the components given the limited stochastic duration of the break and the minimum system reliability level required to complete the next mission. The fundamental concepts and relevant parameters of this decision-making problem are developed and discussed. Numerical experiments are provided to demonstrate the added value of solving this selective maintenance problem as a stochastic optimisation programme. [ABSTRACT FROM AUTHOR]

• Novel methodology for variable flow CO 2 pipeline design optimisation. • Utilises logit-normal distribution to describe variable flow scenarios. • Methodology is validated for different flow rates and transport lengths. • Oversized pipeline diameter minimises unitised transport cost. • Oversized design yields lower cost compared to buffer tanks. Pipeline transport has emerged as the most cost-effective method for transporting CO 2 onshore. The CO 2 flow rate is a key factor driving transport costs, underscoring the need to understand the impact of flow rate variability in CO 2 pipeline networks on their design and economics. This paper presents an optimal pipeline design methodology for CO 2 pipelines operating under variable flow that considers the probability distribution of CO 2 flow rate and the pipeline length. The results imply that pipelines designed for optimal performance under variable flow rates often demand a higher level of overdesign compared to pipelines intended for steady-state conditions. Decision-makers must balance the trade-offs between pipeline oversizing and installing multiple pressure boosting stations, especially applicable to large transportation distances and projects of extended duration. The examination of different approaches to pipeline design reveals that a variable flow pipeline design based on mean flow rate is not recommended, because it is incapable of handling the maximum flow rate. This drawback is overcome by adopting the variable flow stochastic pipeline design presented in this paper. [ABSTRACT FROM AUTHOR]

Variance reduction is a crucial tool for improving the slow convergence of stochastic gradient descent. Only a few variance-reduced methods, however, have yet been shown to directly benefit from Nesterov's acceleration techniques to match the convergence rates of accelerated gradient methods. Such approaches rely on "negative momentum", a technique for further variance reduction that is generally specific to the SVRG gradient estimator. In this work, we show for the first time that negative momentum is unnecessary for acceleration and develop a universal acceleration framework that allows all popular variance-reduced methods to achieve accelerated convergence rates. The constants appearing in these rates, including their dependence on the number of functions n, scale with the mean-squared-error and bias of the gradient estimator. In a series of numerical experiments, we demonstrate that versions of SAGA, SVRG, SARAH, and SARGE using our framework significantly outperform non-accelerated versions and compare favourably with algorithms using negative momentum. [ABSTRACT FROM AUTHOR]

We propose a novel method for training deep neural networks that are capable of interpolation, that is, driving the empirical loss to zero. At each iteration, our method constructs a stochastic approximation of the learning objective. The approximation, known as a bundle, is a pointwise maximum of linear functions. Our bundle contains a constant function that lower bounds the empirical loss. This enables us to compute an automatic adaptive learning rate, thereby providing an accurate solution. In addition, our bundle includes linear approximations computed at the current iterate and other linear estimates of the DNN parameters. The use of these additional approximations makes our method significantly more robust to its hyperparameters. Based on its desirable empirical properties, we term our method Bundle Optimisation for Robust and Accurate Training (BORAT). In order to operationalise BORAT, we design a novel algorithm for optimising the bundle approximation efficiently at each iteration. We establish the theoretical convergence of BORAT in both convex and non-convex settings. Using standard publicly available data sets, we provide a thorough comparison of BORAT to other single hyperparameter optimisation algorithms. Our experiments demonstrate BORAT matches the state-of-the-art generalisation performance for these methods and is the most robust. [ABSTRACT FROM AUTHOR]

At the heart of financial mathematics lie stochastic optimisation problems. Traditional approaches to solving such problems, while applicable to broad classes of models, require specifying a model to complete the analysis and obtain implementable results. Even then, the curse of dimensionality challenges the viability of conventional methods to settings of practical relevance. In contrast, machine learning, and reinforcement learning (RL) particularly, promises to learn from data and overcome the curse of dimensionality simultaneously. This article touches on several approaches in the extant literature that are well positioned to merge our traditional techniques with RL. [ABSTRACT FROM AUTHOR]

Copyright of Transformations in Business & Economics is the property of Vilnius University and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Spatially separated locations may differ greatly with respect to their electricity demand, available space, and local weather conditions. Thus, the regions that are best suited to operating wind turbines are often not those where electricity is demanded the most. Optimally, renewable generation facilities are constructed where the maximum generation can be expected. With transmission lines limited in capacity though, it might be economically rational to install renewable power sources in geographically less favourable locations. In this paper, a stochastic bilevel optimisation is developed as a mixed-integer linear programme to find the socially optimal investment decisions for generation expansion in a multi-node system with transmission constraints under an emissions reduction policy. The geographic heterogeneity is captured by using differently skewed distributions as a basis for scenario generation for wind speeds as well as different opportunities to install generation facilities at each node. The results reinforce that binding transmission constraints can greatly decrease total economic and emissions efficiency, implying additional incentives to enhance transmission capacity between the optimal supplier locations and large demand centres.

This paper develops and analyses a stochastic optimisation problem with a service level constraint for generating a sequentially optimal plan of production, maintenance and delivery activities in a deteriorating manufacturing system. Stochastic demand along with product returns are both assumed the latter of which allows for restocking products returned by the customer which are still new and thus in saleable condition. A constrained production/maintenance/delivery problem with service level, stochastic demand, delivery time, failure rate and product returned is formulated based on quadratic model. This quadratic formulation is adapted to provide an inventory, delivery, production and maintenance policies. The objective of this paper is to study the delivery time influence on the planning of the production, maintenance and delivery activities. Finally, we present simulation results to illustrate the exploitation of the proposed approach. [ABSTRACT FROM PUBLISHER]

Stochastic gradient descent is an optimisation method that combines classical gradient descent with random subsampling within the target functional. In this work, we introduce the stochastic gradient process as a continuous-time representation of stochastic gradient descent. The stochastic gradient process is a dynamical system that is coupled with a continuous-time Markov process living on a finite state space. The dynamical system—a gradient flow—represents the gradient descent part, the process on the finite state space represents the random subsampling. Processes of this type are, for instance, used to model clonal populations in fluctuating environments. After introducing it, we study theoretical properties of the stochastic gradient process: We show that it converges weakly to the gradient flow with respect to the full target function, as the learning rate approaches zero. We give conditions under which the stochastic gradient process with constant learning rate is exponentially ergodic in the Wasserstein sense. Then we study the case, where the learning rate goes to zero sufficiently slowly and the single target functions are strongly convex. In this case, the process converges weakly to the point mass concentrated in the global minimum of the full target function; indicating consistency of the method. We conclude after a discussion of discretisation strategies for the stochastic gradient process and numerical experiments. [ABSTRACT FROM AUTHOR]

This paper proposes an online constrained extremum-seeking approach for an unknown convex function with unknown constraints within a class of uncertain dynamical systems with an available output disturbed by a stochastic noise. It is assumed that the objective function along with the constraints are available for measurement. The main contribution of the paper is the formulation of the problem using the penalty method and the development of an extremum seeking algorithm based on a modified synchronous detection method for computing a stochastic gradient descent procedure. In order to reject the undesirable uncertainties and perturbations of the dynamic plant from the beginning of the process, we employ the standard deterministic Integral Sliding Mode Control transforming the initial dynamic plant to static one. Then, we apply the gradient decedent technique. We consider time varying parameters of the suggested procedure for compensating the unknown dynamics. To validate the exposition, we perform a numerical example simulation. [ABSTRACT FROM AUTHOR]

There is a growing interest for the design and operation of reverse supply chain systems due to the cost and the legislation issues. In this paper, we address the disassembly, refurbishing and production operations in a reverse supply chain setting for modular products such as computers and mobile phones considering the uncertainties in this system, which are the return amounts of the used products and demand for final products. We develop a large-scale mixed integer programming model in order to capture all characteristics of this system, and use two-stage stochastic optimisation and robust optimisation approaches to analyse the system behaviour. In the first stage, we focus on the strategic decisions about the capacities at disassembly and refurbishing sites considering different scenarios regarding the uncertainties in the system. In the second stage, we analyse the operational decisions such as production, inventory and disposal rates. We observe through our extensive numerical analysis that the randomness of demand and return values effect the performance of the system substantially and the uncertainty of the return amounts of used products is much more important than the uncertainty of demand in this system. [ABSTRACT FROM PUBLISHER]