关键词: 量子遗传算法, 小生境, 分布估计算法, 扩展紧致遗传算法

Abstract:

The traditional quantum genetic algorithm is slow in convergence speed and is easy to be trapped into local optimum. In order to overcome the above problems and enhance the convergence performance of the quantum genetic algorithm, we propose a novel niche estimation of distribution quantum genetic algorithm integrated with the fitness sharing function method .The quantum chromosome can be rotated in two steps in every subpopulation: the first step is the multigranularity mechanism and the second step is the marginal product model (MPM) rotation. The quantum chromosome crossover based on the MPM can enhance the diversity of the population and avoid the loss of good models. The traits of convergence are also analyzed in the paper, and the entropy convergence criteria are proposed. Functional simulation results show that the proposed algorithm outperforms other traditional algorithms.

Key words: quantum genetic algorithm;niche；estimation of distribution algorithm；extended compact genetic algorithm