可变种群规模的遗传算法

Genetic Algorithm with Variable Population Size

通过模拟人类进化过程中人口数量的增长规律,提出了一种可变种群规模的遗传算法(VPGA)。VPGA不需要改变现有遗传算法的算子,仅仅通过对现有算法框架的重构,从而获得比现有遗传算法更优的解,而花费的计算代价更小。我们的实验结果表明使用离散逻辑斯蒂模型来控制种群规模的VPGA能够比其他从微观算子上改进的遗传算法更加高效省时。

By simulating the population increasemnet in the human evolution, a genetic algorithm with variable population size （VPGA） was proposed. Without changing the GA operations, VPGA can obtain better solutions with less time cost by modifying the structure of traditional GA. Experimental results show that VPGA using logistic model population size is more efficient and requires less computation time than other modified GAs which only improve the GA operations.