关键词: 计算机工程, 遗传算法, 区间算法, 全局优化, 多峰值

Abstract: To overcome the disadvantage of high computation cost in traditional interval optimization algorithms for high dimensional problems, an interval-genetic algorithm was proposed. This algorithm combines the interval algorithm and a genetic algorithm. It employed the interval algorithm to bound the search domains of the genetic algorithm, and adopted a reject index to make the intervals containing the global optimum being more easily selected. Moreover, the algorithm used an upper bound of the global optimum provided by the genetic algorithm to discard the intervals not containing the global optimal solution. Simulation results on some multipeak global optimizations show that the efficiency of the proposed algorithm is higher than traditional interval optimization algorithms and this advantage becomes more significant in solving high dimensional optimizations.

Key words: computer engineering, genetic algorithm, interval algorithm, global optimization, multi-peak