A memetic algorithm for hardware software partitioning

Abstract

Hardware/software partitioning is one of the crucial steps of codesign systems. Most formulations of hardware/software partitioning problem are NP-hard. In this paper, we investigate the application of the memetic algorithm for solving the hareware/software partitioning problem on an NP-hard model. The memetic algorithm uses a local search procedure, which is based on simulated annealing, to improve initial individuals. Experimental tests are done on some instances taken from the literature. The results and comparisons show that the proposed algorithm is efficient. © 2012 Springer-Verlag GmbH.