多宇宙并行量子衍生遗传算法研究

Research of Multi-universe Parallel Quantum-inspired Genetic Algorithm

将量子的多宇宙特性和遗传算法相结合,提出了多宇宙并行量子衍生遗传算法。算法中将种群分成若干个独立的子群体,称为宇宙。给出了不同宇宙数量下的并行拓朴结构,提出了宇宙内采用量子旋转门演化和量子变异,宇宙间采用移民和量子交叉的两种信息交互方式,能有效克服早熟收敛现象,使得搜索效率和搜索能力得到更进一步提高。典型函数优化实验验证了该文所提算法的有效性。

A novel Multi-universe Parallel Quantum-inspired Genetic Algorithm(MPQGA)is put forward based on the combination of quantum's multi-universe property and genetic algorithm.In the algorithm,the colony is divided into sev-eral independent sub-colonies,called universes.Different topological structures according to the number of universes are provided.Two information exchange methods are put forward,which are quantum rotate gate evolutionary and quantum mutation in universe,the best emigration and the quantum crossover operation among universes.These methods can over-come the pre-maturity effectively and improve the search efficiency and search ability greatly.Typical function optimiza-tion validates the effectiveness of MPQGA.