Optimization

Optimisation problems in cellular communication systems and terrestrial transportation involve the distribution of a set of entities over an area (the demand) and a set of physical systems for which an optimal configuration relative to the demand constitutes the solution to the problem. To address such problems, we are interested in bio-inspired and natural heuristics and metaheuristics which exhibit a high potential for distribution and parallel execution, as well as flexibility, modularity and simplicity. Two types of approaches are investigated and applied to optimisation problems.

Adaptive meshing in evolutionary approach

We propose the concept of adaptive meshing to model graph transportation networks, vehicle routes or cellular honeycomb meshes for communication. The solution of the problem is a planar mesh which continuously distorts and adapts locally according to the demand. A population based metaheuristic encapsulates local search operators and neural network self-organizing maps in order to implement the concept and apply it to optimisation problems. The approach has been applied to cellular network dimensioning, to the optimisation of enterprise mobility plans, and to several (static and dynamic) combined vehicle routing and clustering problems.

Coalition-based metaheuristic

We also study how concepts from multiagent systems’ domain may contribute to the design and hybridisation of new metaheuristics. Basing upon an organisational view of metaheuristics, we propose a coalition based metaheuristic framework incorporating evolutionary algorithm and distributed artificial intelligence concepts. Search agents cooperate inside a coalition in a decentralized and asynchronous way. In addition to classical evolutionary and mimetic algorithms, these agents have decision, learning and mimetic abilities. The approach has been applied to vehicle routing and facility location problems.

This page was last modified on 29 July 2010, at 12:09. This page has been accessed 7,808 times.
Copyright 2010-2018 © Laboratoire Électronique Informatique Image - Université Bourgogne Franche-Comté - Privacy policy