Parallel Evolutionary Algorithms for Multiobjective Placement Problem

Non-deterministic iterative heuristics such as Tabu Search (TS), Simulated Evolution (SimE), Simulated Annealing (SA), and Genetic Algorithms (GA) are being widely adopted to solve a range of hard optimization problems [1]. This interest is attributed to their generality, ease of implementation, and their ability to deliver high quality results. However, depending on the size of the problem, such heuristics may have very large runtime requirements. One practical approach to speeding up their execution is parallelization. This is all the more true for multi-objective cell placement, where the need to optimize conflicting objectives (interconnect wire-length, power dissipation, and timing performance) adds another level of difficulty [2]. In this paper, we present parallelization of TS and SimE, for multiobjective VLSI standard cell placement problem. Profile analysis of sequential code is conducted to assist in selecting and engineering earlier proposed strategies. Fuzzy logic is used to integrate the costs of multi-objectives. The TS implementation is a based on a synchronous candidate list partitioning model. The implementation of parallel SimE is based on random distribution of rows to processors [3, 4]. For comparison purposes, a parallel genetic algorithm (GA) based on the island model [5, 6], and a parallel SA based on the Asynchronous Multiple-Markov Chain Model [7] were implemented. Results of experiments on ISCAS-85/89 benchmark circuits are presented, with solution quality and speedup used as metrics for the comparative/relative evaluation of the heuristics.