On the use of genetic algorithms to solve location problems

This paper seeks to evaluate the performance of genetic algorithms (GA) as an alternative procedure for generating optimal or near-optimal solutions for location problems. The specific problems considered are the uncapacitated and capacitated fixed charge problems, the maximum covering problem, and competitive location models. We compare the performance of the GA-based heuristics developed against well-known heuristics from the literature, using a test base of publicily available data sets. Genetic algorithms are a potentially powerful tool for solving large-scale combinatorial optimization problems. This paper explores the use of this category of algorithms for solving a wide class of location problems. The purpose is not "prove" that these algorithms are superior to procedures currently utilized to solve location problems, but rather to identify circumstances where such methods can be useful and viable as an alternative/superior heuristic solution methos.