A PSEUDOGENETIC ALGORITHM FOR MLP DESIGN BASED UPON THE SCHMIDT PROCEDURE

摘要

Based in the natural law of selection a pseudogenetic algorithm is defined in this paper for designing Multilayer Perceptron (MLP). The algorithm is based in two operations; combine and prune. A non-heuristic pruning algorithm is defined from the Schmidt procedure, where orthonormal systems of basis functions are found, ordered, pruned, and mapped back to the original combined network at each iteration of the algorithm. Starting from a population of networks, then during each iteration the algorithm combines an input network and selects the most useful hidden neurons for generating a new network. A complete definition of the algorithm is presented and numerical results are given.