Lamarckian genetic algorithms applied to the aggregation of preferences

The problem that we deal with consists in aggregating a set of individual preferences into a collective linear order summarizing the initial set as accurately as possible. As this problem is NP-hard, we apply heuristics to find good approximate solutions. More precisely, we design a Lamarckian genetic algorithm by hybridizing some meta-heuristics (based on the simulated annealing method or the noising method) with a genetic algorithm. For the problems that we studied, the experiments show that such a hybridization brings improvements to these already good methods.