AN EVOLUTIONARY NEURAL-NETWORK APPROACH FOR MODULE ORIENTATION PROBLEMS

A novel neural network approach called ''Evolutionary Neural Network ( ENN)'' is presented for the module orientation problem. The goal of th is NP-complete problem is to minimize the total wire length by flippin g circuit modules with respect to their vertical and/or horizontal axe s of symmetry, In order to achieve high quality VLSI systems, it is st rongly desired to solve the problem as quickly as possible in the desi gn cycle. Based on the concept of the genetic algorithm, the evolution ary initialization scheme on neuron states is introduced so as to prov ide a high quality solution within a very short time, The performance of ENN is compared with three heuristic algorithms through simulations on 20 examples with up to 500 modules. The results show that ENN can find the best solutions in the shortest time.