David Meignan

Simulation and Evaluation of Urban Bus Networks Using a Multiagent Approach

Abstract Evolution of public road transportation systems requires analysis and planning tools to improve service quality. A wide range of road transportation simulation tools exist with a variety of applications in planning, training and demonstration. However, few simulation models take into account traveler behaviors and vehicle operation specific to public transportation. We present in this paper a bus-network simulation tools which include these specificities and allows to analyze and evaluate a bus-network at diverse space and time scales. We adopt a multiagent approach to describe the global system operation as behaviors of numerous autonomous entities such as buses and travelers.

Coalition-based metaheuristic: a self-adaptive metaheuristic using reinforcement learning and mimetism

We present a self-adaptive and distributed metaheuristic called Coalition-Based Metaheuristic (CBM). This method is based on the Agent Metaheuristic Framework (AMF) and hyper-heuristic approach. In CBM, several agents, grouped in a coalition, concurrently explore the search space of a given problem instance. Each agent modifies a solution with a set of operators. The selection of these operators is determined by heuristic rules dynamically adapted by individual and collective learning mechanisms. The intention of this study is to exploit AMF and hyper-heuristic approaches to conceive an efficient, flexible and modular metaheuristic. AMF provides a generic model of metaheuristic that encourages modularity, and hyper-heuristic approach gives some guidelines to design flexible search methods. The performance of CBM is assessed by computational experiments on the vehicle routing problem.

A Review and Taxonomy of Interactive Optimization Methods in Operations Research

This article presents a review and a classification of interactive optimization methods. These interactive methods are used for solving optimization problems. The interaction with an end user or decision maker aims at improving the efficiency of the optimization procedure, enriching the optimization model, or informing the user regarding the solutions proposed by the optimization system. First, we present the challenges of using optimization methods as a tool for supporting decision making, and we justify the integration of the user in the optimization process. This integration is generally achieved via a dynamic interaction between the user and the system. Next, the different classes of interactive optimization approaches are presented. This detailed review includes trial and error, interactive reoptimization, interactive multiobjective optimization, interactive evolutionary algorithms, human-guided search, and other approaches that are less well covered in the research literature. On the basis of this review, we propose a classification that aims to better describe and compare interaction mechanisms. This classification offers two complementary views on interactive optimization methods. The first perspective focuses on the user’s contribution to the optimization process, and the second concerns the components of interactive optimization systems. Finally, on the basis of this review and classification, we identify some open issues and potential perspectives for interactive optimization methods.

A Coalition-Based Metaheuristic for the vehicle routing problem

This paper presents a population based Metaheuristic adopting the metaphor of social autonomous agents. In this context, agents cooperate and self-adapt in order to collectively solve a given optimization problem. From an evolutionary computation point of view, mechanisms driving the search consist of combining intensification operators and diversification operators, such as local search and mutation or recombination. The multiagent paradigm mainly focuses on the adaptive capabilities of individual agents evolving in a context of decentralized control and asynchronous communication. In the proposed metaheuristic, the agentpsilas behavior is guided by a decision process for the operatorspsilachoice which is dynamically adapted during the search using reinforcement learning and mimetism learning between agents. The approach is called Coalition-Based Metaheuristic (CBM) to refer to the strong autonomy conferred to the agents. This approach is applied to the Vehicle Routing Problem to emphasize the performance of learning and cooperation mechanisms.

A heuristic approach to schedule reoptimization in the context of interactive optimization

Optimization models used in planning and scheduling systems are not exempt from inaccuracies. These optimization systems often require an expert to assess solutions and to adjust them before taking decisions. However, adjusting a solution computed by an optimization procedure is difficult, especially because of the cascading effect. A small modification in a candidate solution may require to modify a large part of the solution. This obstacle to the adjustment of a solution can be overcome by interactive reoptimization. In this paper we analyze the impact of the cascading effect on a shift-scheduling problem and propose an efficient heuristic approach for reoptimizing solutions. The proposed approach is a local-search metaheuristic that has been adapted to the reoptimization. This approach is evaluated on a set of problem instances on which additional preferences are generated to simulate desired adjustments of a decision maker. Experimental results indicate that, even with a small perturbation, the cascading effect is manifest and cannot be efficiently tackled by applying recovery actions. Moreover, results show that the proposed reoptimization method provides significant cost gains within a short time while keeping a level of simplicity and modularity adequate for an implementation in a decision support system.

An interactive heuristic approach for the P-forest problem

In this paper, we propose and compare two complementary heuristic approaches for solving the P-forest problem. The first one is a Greedy Randomized Adaptive Search Procedure (GRASP), and the second one is an interactive heuristic approach. Contrary to the GRASP, which is a fully automated approach, in the interactive heuristic the user contributes in a cooperative manner to the optimization process. The objective is to exploit the problem-domain expertise of the user in order to generate more realistic solutions that integrate aspects not captured by the objective function. These heuristics were implemented on a decision support system for solving a P-forest problem in the domain of forestry. We present experimental results on real problem instances of access road networks design. A comparison between manual planning and the two heuristics shows clear advantages for using the proposed interactive approach.

A Self-Organizing and Holonic Model for Optimization in Multi-Level Location Problems

Multi-Level Location Problems are generally considered as complex. To deal with these problems we propose an approach based on Holonic MultiAgent Systems (HMAS). HMAS have already proven to be a convenient way to engineer complex systems. This approach was merged with Artificial Potential Fields (APF) mechanims. The solution is obtained simultaneously with the holarchy. The holarchy is thus used to exploit and control the emergence of the solution. This solution is then evaluated to check its relevance according to global objectives represented thanks to a fitness function. This model was efficiently applied to a multi-level distribution system.

Un framework organisationnel et multi-agent pour la conception de métaheuristiques

This paper presents the AMF framework that aims at supporting the design of metaheuristics using an organizational and multiagent approach. It aims at comparing and analyzing existing algorithms, and also facilitating the design of hybrid and new heuristics. This framework is based on an organizational model which describes the metaheuristics in terms of roles and interactions. In addition, we provide some methodological guidelines which favor the design of distributed and auto-adaptive metaheuristics.