Hierarchical Neural Techniques for Location Problems

E. Domínguez and J. Muñoz (Spain)


Combinatorial optimization, p-median problem, recurrent neural networks, location.


There exist several neural network techniques for solving NP-hard combinatorial optimization problems. Hopfield networks and self-organizing maps are the two main neu ral approaches studied. Criticism of these approaches in cludes the tendency of the Hopfield network to produce in feasible solutions and the lack of generalization of the self organizing approaches. In this paper we propose a compet itive recurrent neural model for solving location problems which enables feasibility of the solutions and improved so lution quality through escape from local minima. Based on this neural model, three new hierarchical neural tech niques are proposed for solving location problems. The effectiveness and efficiency of the three neural techniques under varying problem sizes are analyzed. The results in dicate that the best technique depend on the scale of the problem.

Important Links:

Go Back