Résolution des problèmes difficiles par optimisation distribuée

Abstract

The hybridization and parallelism of metaheuristics are used successfully to solve large difficult problems. Parallel hybrid metaheuristics require enormous computing power. The grids are a distributed environment that can provide that power. The grids are characterized by their diverse, volatility of its resources and time latency. These features must be respected when designing methods and parallel hybrids. This work presents a result of hybridization between the particle swarm optimization and local search iterated. This method required to be adapted to grid computing environment