Neural networks for the set covering problem: an application to thetest vector compaction

摘要

A new neural network architecture is proposed and evaluated fornthe set covering problem. The network is designed by using thenconnections between nonlinear and integer programming problems. Thisnconnection is based on the concavity and penalty function methods. Thengeneral objective function obtained, which combines the objectivenfunction and constraints, is fixed as the energy of the system. Thennetwork obtained was then applied to a practical problem in VLSInsystems. The simulation results for the network show that the system cannconverge rapidly within a few neural time constants even for large scalenproblems