Search

Aircraft routing and maintenance planning still represents a challenging research topic. Aircraft routing, or aircraft tail assignment, optimises fleet allocation to predefined routes. Maintenance is optimized by accordingly making decisions based on aircraft preventive or condition-based needs. These two problems can lead to separate optimization processes. Nevertheless, a combined optimization problem should produce better results, as routes should impact maintenance costsand vice-versa. One issue is how to scale the problem such that many aircraft could fly at the same time from one airport to another and make maintenance decisions in an uncertain environment due to health indicators dynamics. In this work, we propose a unified model for this problem based on Markov Decision Processes (MDP). Two limiting factors impact our search of solution: the number of maintenance actions to be considered in planning that increases with the number of aircraft, as well as, the number of possible allocations that in turn is affected by the number of aircraft and airports. We propose a hybrid planning solution which decomposes the problem in such way it performs aircraft tail assignment optimisation combined with maintenance policy. The maintenance policies are computed for several independent MDPs (e.g. one per aircraft) in a robust manner by considering several possible aircraft locations. Then, a solution for the aircraft tail allocation problem is obtained by using aircraft maintenance policies. Finally, we empirically show that this method scales better than a unified (and possibly huge) MDP approach.

The Military Flight and Maintenance Planning (MFMP) Problem assigns missions and schedules maintenance operations (checks) for military aircraft. It has been studied in short, medium and long-term planning horizons ([1, 2, 5, 4]) and it has been proved NP-Hard in its long-term variant by [3]. The present model is an alternative formulation to that of [3] where mission assignments and maintenance cycles are modeled as start-stop assignment.

International audience; Aircraft equipment health monitoring system plays a promising role for airlines operation cost reduction, as it can be exploited to perform predictive maintenance. In this vein, a hierarchical sequential decision making model is proposed to plan predictive maintenance. It combines linear optimization for routing assignment and MDP planning to handle maintenance actions based on the stochastic evolution of health indicators. This entangled model should reduce planning time while ensuring a cost-efficient policy.

International audience; This abstract is focused on the Periodically Aggregated Resource-Constrained Project Scheduling Problem (PARCPSP) ), that can be seen as a continuous-time variant of a restricted Resource-Constrained Project Scheduling Problem with partially renewable resources (RCPSP/π) (Böttcher et. al. 1999). The purpose of this work is to compare an existing compact formulation with a new extended formulation.

International audience

International audience; A Flight and Maintenance Planning problem is considered for military aircraft. A novel Mixed-Integer Linear Programming model is developed for scheduling maintenance operations along with the assignment of regular, pre-scheduled, missions. This model takes into account the specific requirements of a military environment and its operational goals. On the contrary to previous studies, the developed model is applicable for the management of a heterogenic fleet. Different combinations of objective functions are discussed for multi-objective optimization and the impact of their use is analyzed on a dataset provided by French Army Forces. The complexity of the obtained problem formulations is studied; new procedures are developed in order to decrease the computational time required to construct efficient Pareto-fronts of alternative solutions for real-life sized instances. The results obtained in computational tests show the efficiency of the developed optimization approach.

The Military Flight and Maintenance Planning Problem considered here aims to assign missions and maintenance tasks to military aircraft. It is a variant of the Flight and Maintenance Planning problem where flights are not modeled geographically since a round-trip to the base is assumed for each flight. It also includes different objectives and constraints.

The Flight and Maintenance Planning (FMP) problem deals with the scheduling of maintenance operations and the assignment of flight activities to a fleet of aircraft. It has two main variants: civil (or commercial) and military. We study the latter, lesser known variant. Previous work include short and medium term planning.

A specific Flight and Maintenance Planning problem is presented. In this problem, preventive maintenance operations are scheduled for military aircraft along with the assignment of regular missions. The quality of the solutions is measured via a bi-objective function that smooths both maintenance operations and aircraft unavailability among time periods. A real world dataset provided by the French Air Force is used to test an exact optimization approach based on a Mixed Integer Programming model. Mono-objective computations are solved to optimality for medium size instances, allowing to compute exactly the Pareto frontier of the bi-objective optimization problems. The tests show that these two objectives do not lead to the same optimal solution, but very good compromise solutions can be found thanks to the bi-objective optimization.

Large-scale engineering and construction projects are subject to a great level of uncertainty which lead planners to use time buffers and add contingencies to the estimated budget. However, the size of the buffers and the contingency amounts are usually arbitrarily established and projects still encounter severe time and cost overruns. In this paper, a robust planning approach for tactical planning of large-scale engineering and construction projects is proposed. The approach relies on a simple resource buffering strategy applied to the aggregate periods. An extensive simulation-based experiment was conducted to test the robustness and performance of the proposed approach. Results show that the proposed buffering strategy can considerably reduce project cost variations and can provide comparable performance results with those obtained using a disaggregated approach, especially on instances characterized by a large number of resource groups.

International audience; In this paper we propose to partially relax the resource constraints in order to get lower bounds of the RCPSP. To this aim, by analogy with tactical project planning, we consider the resource constraints on aggregated periods while the precedence constraints still hold. Hence, according to period duration, the resource constraints will be more or less strong. An iterative bounding scheme is presented and some results are given. 1 A time/resource relaxation of the Resource-Constrained Project Scheduling Problem Given a set of activities V = {0,. .. , n + 1} linked by precedence constraints E, and such that each activity i ∈ V requires an amount b i,k ≥ 0 on each resource k of a set of resources R, the Resource-Constrained Project Scheduling Problem (RCPSP) can be defined as the problem of finding a start time S i for each activity i ∈ V while satisfying the precedence constraints E and not exceeding the resource availabilities B k , k ∈ R, such that the makespan, given by the difference between the start time of the dummy end activity n + 1 and the dummy start activity 0, is minimized. We consider a relaxation of the RPCP (P ∆) for any ∆ ∈ R * +. (P ∆) min S n+1 − S 0 (1) S j ≥ S i + p i (i, j) ∈ E (2) i∈V p i,t b i,k ≤ ∆.B k t ∈ T, k ∈ R (3) p i,t = |[t∆, (t + 1)∆] ∩ [S i , S i + p i ]| i, ∈ A, t ∈ T (4) S i ≥ 0 i ∈ A (5) where T = {0,. .. , |T | − 1} is a set of periods t of length ∆ and p i,t denote the length of the time interval under which activity i is executed in period t. We assume that (|T | − 1)∆ is an upper bound on the makespan. To obtain a correct formulation for the RCPSP, one should take ∆ = 1 and enforce integrity on the start times, i.e. replace constraints (5) by S i ≥ 0 and integer , ∀i ∈ A. Obviously (P ∆) is a relaxation of the RCPSP (even for ∆ = 1) as it can be seen that any schedule S feasible for the RCPSP, is also feasible for (P ∆), for any ∆ ∈ R * +. Furthermore, in the figure below with three activities having requirements (3,4,3) on a single resource of availability 4 a makespan of 6 is obtained thanks to non integer start times in the optimal solution of (P 2). Note that besides being a valid relaxation of the RCPSP, the problem (P ∆) can be used in a project planning context, when resource consumption have to be aggregated over several periods in an upper decision level while precedence constraints between activities have

International audience; We consider a project scheduling problem in a context such that temporal aspects (start and completion of activities, precedence between activities) are handled precisely, while resource consumption is evaluated more roughly over aggregated periods. Hence, the Periodically aggregated Resource-Constrained Project Scheduling Problem (PARCPSP) is well-suited for an intermediate planning/scheduling decision level, in particular with human resources or energetic resources. This problem has been modelled by means of a Mixed Integer Linear Programming (MILP) formulation, based on a mixed time framework. In this extended abstract, we present an alternative way to manage both continuous and discrete temporal representations simultaneously, in order to derive a new MILP formulation for a purpose of theoretical and computational comparison.

Le chevauchement d'activités est une pratique courante pour réduire la durée des projets d'ingénierie. Cela consiste à autoriser le début d'une activité avant la fin de son prédécesseur, sur la base d'informations préliminaires. Cependant, le chevauchement peut entraîner du travail supplémentaire et des modifications suite à la communication d'informations complémentaires après le début du successeur. Ainsi le gain en temps s'accompagne d'un coût lié au chevauchement.La planification tactique (Rough-Cut Capacity Planning, RCCP), effectuée dans les phases amont des projets, voit les ressources de manière agrégée par périodes (semaine, mois, etc.). Les activités sont groupées en lots de travaux pour lesquels les effectifs peuvent varier d'une période à une autre, modulant ainsi la durée des lots. La planification consiste à définir la réalisation des lots de travaux dans le but de minimiser la durée et/ou le coût global du projet, en respectant les contraintes de précédente entre lots et les contraintes de capacité des ressources par période. Dans ce travail nous proposons un modèle RCCP avec possibilité de chevauchement des lots de travaux. Plusieurs modes de chevauchements entre lots sont possibles, définis par l'avancement (charge de travail réalisée) et la surcharge de travail associée.Le modèle proposé est une extension d'un modèle RCCP existant. Ce modèle de programmation en variables mixtes combine une représentation en temps continu des événements de début et de fin des lots et une représentation en temps discret des considérations de ressources. Des variables binaires permettent de faire le lien entre les événement de début et de fin d'un lot et le temps consacré à un lot dans une période. Cette durée sert à déterminer la charge réalisée pour un lot dans une période, conformément aux limites sur l'affectation des ressources à ce lot. Dans le nouveau modèle, des événements sont ajoutés pour repérer le début possible du chevauchement dans un mode donné, lorsque la charge nécessaire a été réalisée sur le prédécesseur. La surcharge liée au chevauchement est représentée dans le modèle comme une activité supplémentaire, présente tout au long du chevauchement entre le prédécesseur et le successeur.Il n'existe pas à notre connaissance d'autre modèle RCCP proposant plusieurs modes de chevauchement et la prise en compte de la surcharge. Les tests réalisés sur des instances RCCP adaptées mettent en évidence l'intérêt de proposer plusieurs modes de chevauchement entre lots de travaux. Ce levier permet en effet d'améliorer la durée mais aussi dans certains cas le coût par une meilleure répartition de l'utilisation des ressources, malgré la surcharge. Sur les instances testées, les temps de calcul ont augmenté de manière très raisonnable par rapport au cas sans chevauchement.

Dans les grands projets, il est souvent difficile de mesurer les progrès accomplis en raison de la complexité, parce que la réalisation est partagée entre des départements d'une entreprise voire entre des entreprises disséminées de par le monde. La littérature en gestion de projet et en recherche opérationnelle est examinée pour inventorier les différentes techniques applicables. Les outils largement utilisés pour le suivi et la prévision, comme la valeur acquise, graphiques de tendance, ingénierie concourante, peuvent être employés. Cet article se base sur un problème lié à l'industrie pharmaceutique où l'efficacité d'un traitement médical est examinée sur des patients répartis dans un certain nombre de pays. Le nombre des variables impliquées augmente la complexité de ce problème. L’objectif principal est d'étudier l'efficacité d'une solution pour différentes situations d’avancement dans le cadre du projet, de façon à réduire la durée du projet pour un coût acceptable. Nos résultats suggèrent que la possibilité de réaffectation des patients entre pays produit de meilleurs résultats.

International audience; In this paper we present a scheduling problem dealing with energy constraints (typically electrical energy). Mainly we propose an extension of specific resource constraint propagation techniques (known as "energy reasoning") to efficiently prune the search space and then to facilitate its resolution. We also present dominance rules and a branching scheme to solve the problem via tree search. Finally, computational results are provided.

International audience; Progress measurement, in an environment where remaining work can be transferable from one resource to another, is hard to perform due to the ambiguities and complexities involved in the nature of these projects. Thispaper discusses the environs of Transferable Work-load project, the realization of which is transferable among different departments of a same company or in different parts of the world. Project management and operational research literature is reviewed for discovering various planning, management, monitoring, and forecasting techniques applicable to such projects. This paper presents a problem related to pharmaceutical industry which involves the determination of the effectiveness of a given medical treatment, applied to patients in a given number of countries. In this problem patients are first recruited then their treatment starts. The great number of variables involved increases the complexity of this problem. The main objective of the problem is to analyze the effectiveness of a particular solution, for the recruitment of patients, in different situations such that a shorter project duration and in turn a lower project cost can be achieved.

Progress measurement, in an environment where remaining work can be transferable from one resource to another, is hard to perform due to the ambiguities and complexities involved in the nature of these projects. This paper discusses the environs of Transferable Work-load project, the realization of which is transferable among different departments of a same company or in different parts of the world. Project management and operational research literature is reviewed for discovering various planning, management, monitoring, and forecasting techniques applicable to such projects. This paper presents a problem related to pharmaceutical industry which involves the determination of the effectiveness of a given medical treatment, applied to patients in a given number of countries. In this problem patients are first recruited then their treatment starts. The great number of variables involved increases the complexity of this problem. The main objective of the problem is to analyze the effectiveness of a particular solution, for the recruitment of patients, in different situations such that a shorter project duration and in turn a lower project cost can be achieved.

Cet article aborde le problème de l'ordonnancement d'un atelier de type flowshop, lorsque les durées opératoires varient suivant quel opérateur travaille sur quelle machine. La prise en compte de cette notion de performance ajoute l'affectation des opérateurs au problème d’ordonnancement, qui ne se résume plus à une recherche de séquence de passage des travaux sur les machines. Une heuristique basée sur une règle d'affectation "intuitive" généralement utilisée est évaluée, puis une borne inférieure du problème complet est présentée. Cette borne permet de faire apparaître certaines caractéristiques des règles d’affectation.

Parmi les outils les plus importants pour aider à la Gestion de Projets, on trouve des référentiels de connaissances (IPMA, PMI, etc.). L’importance de tels outils est remarquable pour comprendre les particularités des projets et de leur suivi. Ils ne suppriment cependant pas la nécessité de mettre en place un processus méthodologique d’aide à la conception et mise en œuvre de telles particularités. Il existe cependant par ailleurs d’autres méthodes, pas nécessairement orientées vers la gestion de projets, mais qui peuvent contribuer d’une façon importante à lui donner le caractère systématique indispensable dans cette phase qui conditionne largement le succès ou l’échec futurs du projet. De telles méthodes font un traitement systématique particulier des aspects et étapes tels que l’initiation, aspects sociaux, aspects techniques, etc. L’objectif du présent article n’est pas de choisir une méthode particulière, mais d’arriver à trouver et proposer une complémentarité de méthodes existantes, en profitant des apports de chacune d’elles dans le cadre de la conception et mise en ouvre des projets.