Descent with mutations applied to the linear ordering problem

Abstract

We study here the application of the “descent with mutations” metaheuristic to the linear ordering problem. We compare this local search metaheuristic with another very efficient metaheuristic, obtained by the hybridization of a classic simulated annealing with some ingredients coming from the noising methods. The computational experiments on the linear ordering problem show that the descent with mutations provides results which are comparable to the ones given by this improved simulated annealing, or even better, while the descent with mutations is much easier to design and to tune, since there is no parameter to tune (except the CPU time that the user wants to spend to solve his or her problem).