Dissertations / Theses on the topic 'Traitement réparti – Planification'
Create a spot-on reference in APA, MLA, Chicago, Harvard, and other styles
Consult the top 24 dissertations / theses for your research on the topic 'Traitement réparti – Planification.'
Next to every source in the list of references, there is an 'Add to bibliography' button. Press on it, and we will generate automatically the bibliographic reference to the chosen work in the citation style you need: APA, MLA, Harvard, Chicago, Vancouver, etc.
You can also download the full text of the academic publication as pdf and read online its abstract whenever available in the metadata.
Browse dissertations / theses on a wide variety of disciplines and organise your bibliography correctly.
Adnan, Hashmi Muhammad. "Un langage de programmation agent intégrant la planification temporelle et les mécanismes de coordination de plans." Paris 6, 2012. http://www.theses.fr/2012PA066312.
Full textClerc, Xavier. "Planification dans un espace de buts par stratégie de type meilleur d'abord." Grenoble INPG, 2007. http://www.theses.fr/2007INPG0059.
Full textMost of distributed planning systems are based on models which were developped for centralized planning. These models have then been adapted to distribution and its specific contraints. Our goal is at the opposite to design a planning model that considers these constraints as premises. We have developped a planning model that uses a best-first search (as an adaptation of the proof-number search algorithm). We have applied this model to planning over task structures (from multiagent notations) as well as to HTN planning. Ln this latter case, we have shown how a best-first search allows the planner to rapidly gather constraints that can prune branches from the search space. We have also defined plan robustness in order to mitigate the consequences of an agent failure or a resource unavailability
Gaborit, Paul. "Planification distribuée pour la coopération multi-agents." Phd thesis, Université Paul Sabatier - Toulouse III, 1996. http://tel.archives-ouvertes.fr/tel-00142562.
Full textYousfi, Fouad. "Placo : modélisation par workflow et conception d'un système de planification coopérative : application aux unités de soins." Lille 1, 1996. https://pepite-depot.univ-lille.fr/LIBRE/Th_Num/1996/50376-1996-111.pdf.
Full textConforto, nedelmann Déborah. "Vers le passage à l'échelle de l'allocation en ligne multi-agents multi-tâches." Electronic Thesis or Diss., Toulouse, ISAE, 2024. http://www.theses.fr/2024ESAE0049.
Full textThis thesis is set in the context of online multi-agent multi-task allocation, aiming to efficiently coordinate a set of agents (resources) to distribute tasks among them. As an application example, we can cite cases where clients make requests to a service provider who seeks the best distribution of tasks among its set of agents: taxis responding to customer demands or robots ensuring the delivery of parcels.Unlike the offline framework, which assumes full knowledge of the allocation problem to be solved before the start of the allocation, in the online framework, tasks arrive over time and must be allocated dynamically. This online optimization framework presents several challenges. Firstly, the repeated allocation of tasks to agents is an NP-hard problem, whose solution must be found in a limited time, sometimes very short. Secondly, knowledge about the arrival of future tasks (e.g., arrival time and location) is generally modest, making long-term reasoning difficult (i.e., optimizing the positioning of agents for long-term planning). Finally, the size of the problem to be handled in terms of agents and tasks can be substantial, especially in a realistic setting where finding a solution within a limited time can sometimes be compromised.In this context, this thesis proposes various contributions. Firstly, this thesis proposes a proactive approach that anticipates the availability of agents in the near future to achieve efficient coordination (i.e., minimizing the distance traveled by agents and the time they remain idle). This proactive approach has been compared to a classic reactive approach. The results obtained in two benchmark problems, one synthetic and the other based on real data, show that the proactive method achieves better results in terms of costs and the number of tasks allocated to agents compared to a reactive approach, while reducing the idle time of resources. Despite the encouraging results obtained, the proposed method has scalability limitations in our real-data-based benchmark problem. To address this limitation, this thesis proposes a second approach that develops a multi-agent multi-task allocation meta-heuristic called SKATE - Successive Rank-based Task Assignment for Proactive Online Planning, enabling scalability. SKATE allows for handling problems with thousands of agents and tasks, obtaining effective solutions in a limited time. SKATE shows good results in terms of the cost of the solutions found for such a scale of agents and tasks when compared to classical methods in the literature, such as a genetic algorithm or integer linear programming. Thanks to these results, this thesis subsequently considered two extensions to SKATE. The first extension optimizes not only task assignments but also the number of agents to consider. To achieve this, this thesis develops two methods that optimize both the number of agents (resource savings for the service provider) while ensuring user satisfaction (waiting time before task completion). The second extension couples SKATE with verifiable computing tools, allowing agents to verify that task assignment has been correctly performed by the central server and countering cyber-physical attacks that a network of mobile agents could face in hostile environments, for example
Pastorelli, Mario. "Disciplines basées sur la taille pour la planification des jobs dans data-intensif scalable computing systems." Electronic Thesis or Diss., Paris, ENST, 2014. http://www.theses.fr/2014ENST0048.
Full textThe past decade have seen the rise of data-intensive scalable computing (DISC) systems, such as Hadoop, and the consequent demand for scheduling policies to manage their resources, so that they can provide quick response times as well as fairness. Schedulers for DISC systems are usually focused on the fairness, without optimizing the response times. The best practices to overcome this problem include a manual and ad-hoc control of the scheduling policy, which is error-prone and difficult to adapt to changes. In this thesis we focus on size-based scheduling for DISC systems. The main contribution of this work is the Hadoop Fair Sojourn Protocol (HFSP) scheduler, a size-based preemptive scheduler with aging; it provides fairness and achieves reduced response times thanks to its size-based nature. In DISC systems, job sizes are not known a-priori: therefore, HFSP includes a job size estimation module, which computes approximated job sizes and refines these estimations as jobs progress. We show that the impact of estimation errors on the size-based policies is not signifi- cant, under conditions which are verified in a system such as Hadoop. Because of this, and by virtue of being designed around the idea of working with estimated sizes, HFSP is largely tolerant to job size estimation errors. Our experimental results show that, in a real Hadoop deployment and with realistic workloads, HFSP performs better than the built-in scheduling policies, achieving both fairness and small mean response time. Moreover, HFSP maintains its good performance even when the cluster is heavily loaded, by focusing the resources to few selected jobs with the smallest size. HFSP is a preemptive policy: preemption in a DISC system can be implemented with different techniques. Approaches currently available in Hadoop have shortcomings that impact on the system performance. Therefore, we have implemented a new preemption technique, called suspension, that exploits the operating system primitives to implement preemption in a way that guarantees low latency without penalizing low-priority jobs
Zeddini, Besma. "Modèles d'auto-organisation multi-agents pour le problème de transport à la demande." Le Havre, 2009. http://www.theses.fr/2009LEHA0025.
Full textThis PhD thesis is motivated by the proposal of multiagent models for the Dial A Ride Problem with Time Windows (DARPTW). The DARPTW is a highly complex dynamic problem, for which a multiagent design is relevent. Our proposals focus on Self-Organization models in multiagent systems that allows for the consideration of new criteria for the assessment of the proposed systems, which with the strict consideration of the utility of the transport operator. In our work, we propose several multiagent architectures for the implementation of the DARPTW system. After experimentally evaluating the different architectures, we popose algorrithmic improvements of the best architecture. The objective of these improvements is to palliate the drawbacks related to the myopic behavior of insertion heuristics and the sequentiality of their insertion process. On the one side, we relax the constraint on the non-revokation of assignment decisions by allowing vehicles to exchange customers that they have inserted. On the other side, by adopting an extension of the Contract Net Protocol, we propose to Vehicle agents to process several customers in parallel. The third contribution of this PhD thesis is the proposal of two Self-Organization models (spatial and temporal) allowing a better spatial and temporal coverage of the network. A set of experiments validate our proposals. Finally, we implement a platform allowing for the deployment of DARPTW systems
Sumic, Aïdin. "Prise en compte des activités interdépendantes, des durées incertaines et de l'interopérabilité sémantique dans la coordination temporelle des plans muti-agents." Electronic Thesis or Diss., Université de Toulouse (2023-....), 2024. http://www.theses.fr/2024TLSEP114.
Full textIn multi-agent temporal planning problems with uncertain durations, agents must coordinate and synchronize the execution of their tasks (the start and duration of a task). Here, coordination and synchronization are focused on the duration of tasks that an agent controls but are uncontrollable for other agents of the systems. This means a task being performed by one agent of the system has a duration that is decided by the agent executing it (owner) but is uncontrollable for those who observe it. This is due to some qualitative constraints amongst tasks (e.g., precedence relation), such as another agent (observer) needing to wait for the completion of this task to execute its tasks. For such an observer agent, the duration of this task is uncontrollable unless some communication or coordination is made, which can happen before or during the execution of the plan. Nonetheless, this agent needs to find an executable plan for whether such coordination is made. These are the questions this thesis aims to answer: How can we model such temporal coordination problem? How can we negotiate those uncertain durations under the control of one agent to repair a non executable plan? How can agents communicate when they do not share a common representation of time?For the first one, a new multi-agent model is proposed using the semantics of Temporal Networks under Uncertainty to represent temporal constraint as an interval of possible value between two instants, which can be, for instance, the start and end instant of a task. This new Multi-agent Interdependent Simple Temporal Network under Uncertainty (MISTNU) model represents shared tasks as negotiable contracts between agents. This model aims to guarantee the executability of agents 'plans depending on when the duration of the contracts is shared among the agents. If the model is deemed uncontrollable and, hence, not executable, then a repair phase is initiated by the agents that will negotiate the duration of these contracts to ensure the controllability of the model. This thesis proposed multiple solutions to the repair problem of MISTNUs: those that are centralized, assuming a central agent with full authority and observability, and those without such a central agent, resulting in agents independently negotiating the duration of their contracts until a solution is found (if it exists).To ensure interoperability between agents, a formal ontology is provided that gives a common vocabulary for temporal constraints on intervals. This ensures agents can understand each other and properly answer requests
Fayech, Besma. "Régulation des réseaux de transport multimodal : systèmes multi-agents et algorithmes évolutionnistes." Lille 1, 2003. https://pepite-depot.univ-lille.fr/LIBRE/Th_Num/2003/50376-2003-323.pdf.
Full textCorona, Gabriel. "Utilisation de croyances heuristiques pour la planification multi-agent dans le cadre des Dec-POMDP." Electronic Thesis or Diss., Nancy 1, 2011. http://www.theses.fr/2011NAN10026.
Full textIn this thesis, we focus on planning in decentralised sequentialdecision taking in uncertainty. In the centralised case, the MDP andPOMDP frameworks leads to efficient planning algorithms. The Dec-POMDPframework is used to model decentralised problems. This kind ofproblems is in a higher class of complexity than the centralisedproblem. For this reason, until recently, only very small problem could be solved and only for very small horizons. Recently, some heuristic algorithms have been proposed to handle problem of higher size but there is no theoretic proof of the solution quality. In this thesis, we show how to use a heuristic information in the problem, modelled as a probability distribution on the centralised beliefs, to guide the search for a good approximate policy. Using this heuristic information, we formulate each time step of the planning procedure as a combinatorial optimisation problem. This formulation leads to policies of better quality than previously existing approaches
Pastorelli, Mario. "Disciplines basées sur la taille pour la planification des jobs dans data-intensif scalable computing systems." Thesis, Paris, ENST, 2014. http://www.theses.fr/2014ENST0048/document.
Full textThe past decade have seen the rise of data-intensive scalable computing (DISC) systems, such as Hadoop, and the consequent demand for scheduling policies to manage their resources, so that they can provide quick response times as well as fairness. Schedulers for DISC systems are usually focused on the fairness, without optimizing the response times. The best practices to overcome this problem include a manual and ad-hoc control of the scheduling policy, which is error-prone and difficult to adapt to changes. In this thesis we focus on size-based scheduling for DISC systems. The main contribution of this work is the Hadoop Fair Sojourn Protocol (HFSP) scheduler, a size-based preemptive scheduler with aging; it provides fairness and achieves reduced response times thanks to its size-based nature. In DISC systems, job sizes are not known a-priori: therefore, HFSP includes a job size estimation module, which computes approximated job sizes and refines these estimations as jobs progress. We show that the impact of estimation errors on the size-based policies is not signifi- cant, under conditions which are verified in a system such as Hadoop. Because of this, and by virtue of being designed around the idea of working with estimated sizes, HFSP is largely tolerant to job size estimation errors. Our experimental results show that, in a real Hadoop deployment and with realistic workloads, HFSP performs better than the built-in scheduling policies, achieving both fairness and small mean response time. Moreover, HFSP maintains its good performance even when the cluster is heavily loaded, by focusing the resources to few selected jobs with the smallest size. HFSP is a preemptive policy: preemption in a DISC system can be implemented with different techniques. Approaches currently available in Hadoop have shortcomings that impact on the system performance. Therefore, we have implemented a new preemption technique, called suspension, that exploits the operating system primitives to implement preemption in a way that guarantees low latency without penalizing low-priority jobs
Arib, Souhila. "Mécanismes de formation de coalitions d’agents dans les processus de planification." Thesis, Paris 9, 2015. http://www.theses.fr/2015PA090027.
Full textThe work we present, in this thesis, focuses on the coalition formation problem for self-interested agents which plan their activities in multi-agents systems. As a first step, we have proposed, a mechanism that is based on the analysis of the agents' actions in their plans and reasoning about the plans of others. Additionally, we have addressed the problem of coalition formation with dynamic constraints and preferences that agents reveal and communicate to others during their negotiations. Finally, we have refined our coalition formation mechanism allowing a guided search of the coalitions by building a tree of constraints and a tree of coalitions. Each tree is explored by means of the Monte-Carlo algorithm
Arnaud, Jean. "Performance, disponibilité et coût de services Internet adaptatifs." Phd thesis, Grenoble, 2010. http://tel.archives-ouvertes.fr/tel-00529936.
Full textXu, Jin. "Un modèle multi-agent distribué et hybride pour la planification du transport à la demande temps réel." Phd thesis, INSA de Rouen, 2008. http://tel.archives-ouvertes.fr/tel-00558769.
Full textCorona, Gabriel. "Utilisation de croyances heuristiques pour la planification multi-agent dans le cadre des Dec-POMDP." Phd thesis, Université Henri Poincaré - Nancy I, 2011. http://tel.archives-ouvertes.fr/tel-00598689.
Full textBoufaied, Amine. "Contribution à la surveillance distribuée des systèmes à évènements discrets complexes." Toulouse 3, 2003. http://www.theses.fr/2003TOU30234.
Full textZgaya, Hayfa. "Conception et optimisation distribuée d'un système d'information d'aide à la mobilité urbaine : Une approche multi-agent pour la recherche et la composition des services liés au transport." Phd thesis, Ecole Centrale de Lille, 2007. http://tel.archives-ouvertes.fr/tel-00160802.
Full textCes travaux de recherche visent à mettre en œuvre un Système d'Information de Transport Multimodal (SITM) pour optimiser la gestion de flux des requêtes utilisateurs qui peuvent être nombreuses et simultanées. Dans ce cas, le SITM doit pouvoir procéder par des processus de décomposition des requêtes simultanées en un ensemble de tâches indépendantes. Chaque tâche correspond à un service qui peut être proposé par plusieurs fournisseurs d'information, en concurrence, avec différents couts, formats et temps de réponse. Un fournisseur d'information voulant proposer ses services via le SITM, doit d'abord y enregistrer son système d'information, en assumant la responsabilité des aspects juridiques et qualitatifs de ses données. Le SITM est donc lié à un Réseau informatique Etendu et distribué de Transport Multimodal (RETM) qui comporte plusieurs sources d'information hétérogènes des différents services proposés aux utilisateurs de transport.
L'aspect dynamique, distribué et ouvert du problème, nous a conduits à adopter une modélisation multi-agent pour assurer au système une évolution continue et une flexibilité pragmatique. Le système multi-agent proposé s'appuie sur les métaheuristiques pour la recherche et la composition des services; la recherche des services se base sur le paradigme Agent Mobile (AM) utilisant un algorithme d'optimisation dynamique de construction des Plans De Routes (PDR). Cette première étape d'optimisation prépare les parcours des AMs en prenant en considération l'état du RETM. La composition des services utilise les algorithmes évolutionnistes pour optimiser les réponses en termes de coût et de temps, sachant qu'une réponse à une requête utilisateur ne doit pas dépasser un temps maximum autorisé et qu'un utilisateur cherche toujours à avoir le meilleur rapport qualité prix pour les services qu'il demande.
Enfin, le SITM prend en considération les éventuelles perturbations sur le RETM (pannes, goulets d'étranglements, etc.) pour satisfaire les requêtes utilisateurs dans tous les cas de figure. Dans ce contexte, nous avons créé un protocole de négociation entre les agents mobiles et les agents responsables des choix des fournisseurs d'information pour les services demandés, ces agents sont appelés agents Ordonnanceurs. Le protocole proposé dépasse les limites d'une communication agent traditionnelle, ce qui nous a incités à associer au système une ontologie flexible qui permet d'automatiser les différents types d'échanges entre les agents grâce à un vocabulaire approprié.
Les résultats expérimentaux présentés dans cette thèse justifient l'utilisation du paradigme agent mobile dans notre système qui remplace parfaitement bien les paradigmes classiques telle que l'architecture client/serveur. Les simulations présentées montrent différents scénarios de gestion d'un nombre des requêtes simultanées plus ou moins important. En effet, quelque soit le nombre de requêtes utilisateurs formulées pendant un court laps de temps , le système se charge de leur décomposition, de l'identification des services demandés et des fournisseurs d'information susceptibles d'y répondre.
Charrier, Tristan. "Complexité théorique du raisonnement en logique épistémique dynamique et étude d’une approche symbolique." Thesis, Rennes 1, 2018. https://ged.univ-rennes1.fr/nuxeo/site/esupversions/2a4b2a55-42ff-4768-9b9e-677421fef507.
Full textWe study the theoretical complexity of reasoning tasks involving knowledge in multi-agent systems. We consider dynamic epistemic logic (DEL) as a natural way of expressing knowledge, which allows to express nested knowledge of agents and partially observed dynamic actions. We show complexity results for model checking and satisfiability of DEL formulas, and define a symbolic approach for these problems. We also study DEL-based planning and quantification over specific actions: public announcements
Degas, Augustin. "Auto-structuration de trafic temps-réel multi-objectif et multi-critère dans un monde virtuel." Thesis, Toulouse 3, 2020. http://www.theses.fr/2020TOU30058.
Full textIn many areas, simulation is a powerful tool to learn, visualize, and understand the impact of a decision at a given time on the entire system. The area of air navigation is not an exception. Air traffic simulation tools are essential in air traffic management, and must be able to generate a wide variety of realistic scenarios while taking into account different constraints observable by a simulation user, called situations such as traffic density, a flow typology, collisions, a weather event, or any other emerging event. Structuring traffic simulation to achieve realism and different situations is a complex task, given the many objectives and criteria to meet, the diversity of mobile entities and their multiple interactions, and the dynamics of the environment. In the field of air navigation, this complexity is very often managed by humans, whether by the scriptwriter expert who generates the traffic scenario after several hours of trial and error, or by human actors during the simulation that manage the real-time adaptation of the traffic if needed. Traditional resolution approaches have shown their limits to face the complexity of these applications. In this thesis, we propose to solve real-time structuring of a multi-objective traffic simulation by using the AMAS theory (Adaptive Multi-Agents Systems). In these systems, agents pursue local goals and interact in a cooperative manner. By their local interactions, the system is more robust and self-adapt to the dynamics of the environment, allowing the global function to emerge. After several studies, this theory has shown its adequacy to solve complex and dynamic problems. The objective of this work is to model and specialize this theory for the real-tile structuring of multi-objective and multi-criteria traffic simulation. For this purpose, the AGATS agent model with cooperative and local behaviors and interactions has been defined. This model is composed of two sub-models, AGEAS, for structuring the simulation according to a scenario, and CAAMAS, for enabling mobile entities to self-adapt to the scenarios and dynamics of the simulation. The results of the instantiation of these two models for air traffic simulation show the adequacy of the proposed approach for autonomous scenario generation
Toledo, Rojas Francisco. "Analyse d'impacts du lancement de procédés innovants : application des théories multicritères pour une évaluation robuste." Thesis, Université de Lorraine, 2014. http://www.theses.fr/2014LORR0057/document.
Full textThis research belongs to the industrial engineering field, focusing on the first stages of the innovation process (Fuzzy Front-End). In a globalized world, where the technology is near to everyone, enterprises are looking for new technologies to be different from their competitors, offering new and better products / services as a strategy to keep or acquire new competitive advantages. Moreover, innovation is a large and continuous process, where different management models are used. But, because of the complexity of the process, no general model is suitable for situations and projects. Thus, the innovation process involves different stakeholders from the company and its environment. These individuals come from different origins (field of knowledge), have different needs, abilities, importance and work under different restrictions (technological or financial resources, work rules, etc.). Consequently, it is essential for the enterprise to identify all the needs of the parts involved in the products. And, a model is required in order to identify the stakeholder’s needs before starting design tasks. This model aims at more robustness in the new product development process and at the reduction of the failure rate when launching new products. Thus, the objective of this PhD thesis is to propose a new method to analyze the impact of the innovation, allowing companies to identify the dynamic actors' needs, to improve the definition of their new products and to adapt to the possible changes of the market. This tool is part of an innovative projects management methodology, used before the elaboration of the functional specifications of a new product. This locates the research in the first stages of development called Fuzzy Front-End: before the genesis of the product. Fuzzy front ends integrate among others the opportunity / needs identification for innovation to the development of the new product decision taking. The proposed approach for the innovation impact analysis (A.I) consists of two stages: the first one intends to model an innovation ecosystem, integrating the relations, the importance of each stakeholder and their individual necessities in regard to the new product. Through a Dynamic Needs Analysis (A.B.D) it will be possible to get the global needs, which represent the whole system, without considering the disturbances of the environment or the impact of the same innovation. Thus, the second stage of the model, through the generation of innovation impact scenarios (S.I.I), evaluates the impact of the innovation the new product may face at its market launching. Due to a new product emergence changes emerge: actors’ performance, their interrelations, the importance and their own needs about the product [...]
Maalouf, Elie. "A distributed approach for smart production management in cellular manufacturing system for mass customization." Electronic Thesis or Diss., Compiègne, 2022. http://www.theses.fr/2022COMP2689.
Full textThis research tries to answer the following question: How to optimize production planning (including process planning and scheduling) for mass customized products and cellular manufacturing system in an industry 4.0 context, hence in a smart connected factory and supply chain? It proposes a distributed approach for smart production management in cellular manufacturing systems for mass customization. More precisely, it proposes a full approach for manufacturing planning and control for CMS and MC based on dynamic and distributed process/production planning and scheduling. It is based on three decision making levels: 1 - factory level called master planning, integrating real time data from the supply chain; 2 - cell level, called cell planning; 3 - shop floor level called bidding system, dealing with unexpected events. The main research contributions are: 1 - The full distributed approach integrating planning, scheduling, and material handling allocation while considering real time data from the supply chain. 2 - A multi-objective optimization formulation for the master planning problem (factory level). 3 - Two sequence-based resolution approaches implemented on two metaheuristics, Non-dominated Sorting Genetic Algorithm II (NSGAII), and Speed-constrained Multi-Objective Particle Swarm Optimization (SMPSO)
Feki, Mohamed Firas. "Optimisation distribuée pour la recherche des itinéraires multi-opérateurs dans un réseau de transport co-modal." Phd thesis, Ecole Centrale de Lille, 2010. http://tel.archives-ouvertes.fr/tel-00604509.
Full textMejri, Hinda. "Un système d’aide à la régulation d’un réseau de transport multimodal perturbé : réponse au problème de congestion." Thesis, Ecole centrale de Lille, 2012. http://www.theses.fr/2012ECLI0008/document.
Full textTransport networks have been amplified by the increasing number of vehicles and stations and the emergence of new concepts essentially multimodal and intermodal. Thus, the task of managing public transport systems has become very complex and difficult for regulators.To cope with these difficulties, there is the development of systems decision support as an effective solution to traffic control. They can transmit real-time traffic information on transport networks. Our work is based on designing a control system of multimodal transport networks. It may be as an essential tool for effective solutions and real-time to the problem of traffic congestion. It can provide the necessary information to the user in making its decision to move with or without his car. The proposed system is a hybrid between a graph modeling the network and a multi-agent system. This will be supported by an evolutionary approach for generating an optimal control solution. This is justified by the open, distributed and complex network of multimodal transport
Casalino, Matteo Maria. "Approches pour la gestion de configurations de sécurité dans les systèmes d'information distribués." Phd thesis, Université Claude Bernard - Lyon I, 2014. http://tel.archives-ouvertes.fr/tel-01058803.
Full text