The VLSI Circuital Scheme of Generalized Cellular Automata for Parallel Optimization

Abstract

This paper presents a VLSI circuital implementation scheme of generalized cellular automata (GCA) for parallel optimizations. 1 The GCA approach and architecture has been effectively used to solve a class of optimization problems, such as the Travelling Salesmen Problem (TSP) and the Fast Packet Switching Problem (FPSP). In contrast to the Hopfield-type neural network (HNN) and cellular neural network (CNN), the proposed GCA is featured by multigranularity macro-cells and their evolutionary dynamics. The GCA architecture and its hardware implementation scheme has advantages over the HNN and CNN methods in terms of the real-time performance, interconnection complexity, and parameter decision.