Performance Comparison of PSO and Its New Variants in the Context of VLSI Global Routing

Abstract

Substantial reduction of gate delay occurred in recent times owing to radical decrement of transistor size. The interconnect length and delay are accordingly increased owing to the exponential escalation of packaging density with additional transistors being fabri- cated on the same chip area. The function of VLSI routing that seems to be more defying to the scholars, is categorized in global routing and detailed routing phase. In global routing phase, the prevalent method to lessen the wire length for reducing interconnect delay is to adjust the cost of the Steiner tree, devised by the terminal nodes to be interconnected. Nevertheless, Steiner tree problem is a NP-complete problem in classical graph theory where meta-heuristics might impart beneficial elucidations. Particle swarm optimization (PSO) is a robust algorithm concerning VLSI routing field. This chapter is regarding the proposal of a self-adaptive mechanism for monitoring acceler- ation coefficient of PSO and evaluating its functionalities with the existing acceleration coefficient controlled PSO in numerous allocation topologies of terminal nodes within definite VLSI layout. The outcomes of PSO variant with constriction factor in context to VLSI route reduction ability and robustness are also inspected. Additionally, a new effort in adapting the PSO with embracement of genetic algorithm is established.