Abstract:

Floorplanning plays an important role in the physical design of very large scale integrated circuits(VLSI). It is a multiobjective combinatorial optimization and has been proved to be a NPhard problem. To solve this problem,a multiobjective particle swarm optimization (PSO) is proposed. The algorithm adopts sequence pair (SP) representation,thus the particle update formula is modified  by the principle of crossover operator in GA. The concept of ParetoOptimal Solution and elitism preserving strategy are imported. Moreover,a fitness function with phenotype sharing is designed to obtain a more uniformly distributed Pareto front. Experiments on the MCNC benchmarks show the proposed algorithm is feasible and effective.

Key words: floorplanning, multiobjective, particle swarm optimization, sequence pair