Tuned Genetic Algorithms for Finding p-Medians of a Weighted Graph

There are various engineering applications dealing with the prototype problem of nding the best p-medians in a weighted graph. However, the heuristic developments are still of concern due to their complexity. This paper utilizes genetic algorithm as a well-known reliable evolutionary search for such a purpose. Problem formulation is studied, introducing a characteristic graph and specialized genotype representation called \Direct Index Coding«. The genetic operators are also modi ed due to problem requirements, and further tuned using a simulated annealing approach. Such an enhanced evolutionary search tool is then applied to a number of examples to show its e ectiveness regarding the exact results, and to compare eciency between tuned and non-tuned GA.