低偏差蒙特卡罗序列的量子遗传算法  

Quantum Genetic Algorithm Based on Low-discrepancy Monte Carlo Sequences

作　　者：黄山[1] 苏一丹[1] 覃华[1] 蒙祖强[1] 

机构地区：[1]广西大学计算机与电子信息学院,南宁530004

出　　处：《小型微型计算机系统》2017年第2期398-404,共7页Journal of Chinese Computer Systems

基　　金：国家自然科学基金项目(61363027)资助

摘　　要：针对量子遗传算法存在迭代次数多、计算时间长、容易陷入局部极值的问题,提出一种低偏差蒙特卡罗序列量子遗传算法,利用低偏差序列良好的均匀性,实现量子遗传算法探索与利用的平衡.首先,提出新的低偏差序列Hε量子门来更新量子态形式的种群,提高算法探索量子态的能力,减少算法的迭代次数;其次,提出Pareto集邻域搜索,在当前近优解上用低偏差序列在当前解上进行邻域搜索,以寻找更优的解.在5个复杂函数优化问题上验证本文算法,实验结果表明:所提算法的寻优能力较传统量子遗传算法更强,解的质量有两个数量级以上的提高;算法的计算时间和迭代次数亦优于传统量子遗传算法,引入低偏差序列实现量子遗传算法探索与利用的平衡是可行的.To figure out the problems such as the conventional quantum-genetic algorithm is easily prone to fall into local optimum, its much iteration number and its long computing time, a new algorithm named quantum genetic algorithm based on low-discrepancy monte carlo sequences is proposed. This improved algorithm realizes the balance of exploitation and exploration of quantum genetic al- gorithm by utilizing the good uniformity of low-discrepancy sequences. Firstly, a new low-discrepancy sequences H, Q-gate which is used to update the population of quantum states is proposed. It can improve the ability of algorithm exploration quantum state and re- duce the number of iterations of the algorithm. Secondly, Pareto set neighborhood search is put forward. It uses low-discrepancy se- quences for neighborhood search on the current near optimal solution so that the algorithm can look for a more optimal solution. This paper proposed algorithm was validated on five complex function optimization problems. The experimental results show that the searching capability of the proposed algorithm is better than traditional quantum genetic algorithm;the quality of the solution has more than two orders of magnitude increase; this paper proposed algorithm of computing time and the number of iterations is superior to the traditional quantum genetic algorithm. So that quantum genetic algorithm is introduced in the low-discrepancy sequence to achieve the balance of exploration and exploitation is feasible.

关 键 词：量子遗传算法 低偏差序列 低偏差序列Hg量子门 Pareto集邻域搜索 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]