Canadá
This paper examines the performance of improvement search as a function of the quality of the starting solution in the planar (or continuous) p-median problem. We show that using optimal solutions of the analogue discrete p-median problem as the starting solution for heuristic improvement algorithms, as recommended in the literature, can actually lead to inferior performance. That is, good starting solutions obtained in the discrete space with a fraction of the efort can actually be better, a counter-intuitive result that illustrates in a diferent context the less is more principle recently advocated in the literature.
© 2008-2024 Fundación Dialnet · Todos los derechos reservados