Optimisation par colonies de fourmis pour le problème du sac à dos multidimensionnel

We propose an algorithm based on the Ant Colony Optimization (ACO) meta-heuristic for solving the Multidimensional Knapsack Problem (MKP), the goal of which is to find a subset of objects that maximizes a given objective function while satisfying some resource constraints. The proposed ACO algorithm is generic, and we propose three different instantiations, corresponding to three different ways of laying and exploiting pheromone trails. We experimentally compare these three different instanciations. We then experimentally compare our ACO algorithm with two state-of-the-art genetic algorithms, showing that it obtains competitive results. MOTS-CLÉS : Optimisation par colonies de fourmis, Problème du sac à dos multidimensionnel .