Optimization of mixed polarity Reed-Muller expressions based on Whole Annealing Genetic Algorithm

This paper presents a Whole Annealing Genetic Algorithm (WAGA) to obtain a good circuit implementation among mixed polarity Reed-Muller expressions. By combining global searching ability of genetic algorithm and local searching ability of simulated annealing, WAGA could achieve fast convergence. Apart from genetic operators such as crossover operator and mutation operator are used in genetic algorithm stage, it uses annealing operator at annealing stage. At the annealing stage, WAGA forms an intermediate population by selecting 2/3 population from previous generation and 2/3 population from current generation. Annealing operator is then applied to the intermediate population. To achieve an efficient CPU utilization, the calculation of the cost function of WAGA is based on a parallel manner, in which newly generated terms are obtained at one time. The results of tested benchmark show that the algorithm is highly effective for searching the best polarity and achieves 13% improvement on average in terms of CPU time.