A novel method on discrete particle swarm optimization for fixed-outline floorplanning

Abstract

The VLSI fixed-outline fioorplanning which is an important research problem of VLSI design is an NP-hard problem. Particle swarm optimization (PSO) as one of the AI methods, has been proved to be an effective global optimization method. In this paper, an improved discrete PSO combining discussion mechanism and chaos strategy is proposed for the fixed-outline fioorplanning (DCP-FF). A natural sequence representation is adopted for the particle coding, the cost function and related operations are redefined according to the characteristics of the problem. The GSRC benchmarks are employed to test the performance of DCP-FF. The results show that the proposed method is feasible and effective for the fixed-outline floorplanning.