计算机工程与应用 ›› 2008, Vol. 44 ›› Issue (27): 37-40.DOI: 10.3778/j.issn.1002-8331.2008.27.012

• 理论研究 • 上一篇    下一篇

多宇宙并行量子多目标进化算法

李 絮,李智勇,刘松兵,许 波   

  1. 湖南大学 计算机与通信学院,长沙 410082
  • 收稿日期:2008-04-03 修回日期:2008-06-16 出版日期:2008-09-21 发布日期:2008-09-21
  • 通讯作者: 李 絮

Multi-universe parallel quantum-inspired multiobjective evolutionary algorithm

LI Xu,LI Zhi-yong,LIU Song-bing,XU Bo   

  1. College of Computer and Communication,Hunan University,Changsha 410082,China
  • Received:2008-04-03 Revised:2008-06-16 Online:2008-09-21 Published:2008-09-21
  • Contact: LI Xu

摘要: 提出了一种新的基于量子计算的多目标进化算法,即多宇宙并行量子多目标进化算法。算法中将所有的量子个体按给定的拓扑结构分成多个独立子种群,划分为多个宇宙;采用目标个体均匀分配原则和动态调整旋转角机制对各宇宙量子个体进行演化;宇宙之间采用最佳移民操作来交换信息,设计最优个体保留方案以便各宇宙共享全局信息,提高算法的执行效率。该算法用于多目标0/1背包问题的仿真结果表明:新方法能够找到接近Pareto最优前端的更好的解,同时维持解分布的均匀性。

Abstract: This paper proposes a novel multiobjective evolutionary algorithm inspired by quantum computing,which is named Multi-universe Parallel Quantum-inspired Multiobjective Evolutionary Algorithm(MPQMEA).In the algorithm,all individuals are divided into some independent sub-colonies,called universes,according to their definite topological structure.The uniform assignment principle of target individuals and dynamic adjusting rotation angle mechanism are applied to update each universal individuals.Information among the universes is exchanged by adopting the best emigration.To utilize global information,the best reservation scheme is designed for the improvement of search efficiency.Experimental results pertaining to the multiobjective 0/1 knapsack problem show that MPQMEA finds the better solutions close to the Pareto-optimal front while maintaining a uniform spread of nondominated set.