Computer Engineering and Applications ›› 2017, Vol. 53 ›› Issue (7): 54-59.DOI: 10.3778/j.issn.1002-8331.1509-0067

Previous Articles     Next Articles

Quantum genetic algorithm based on universal quantum gates and its applications

LI Sheng1, ZHANG Peilin1, LI Bing2, WU Dinghai1, HU Hao3   

  1. 1.Department Seventh, Ordnance Engineering College, Shijiazhuang 050003, China
    2.Department Fourth, Ordnance Engineering College, Shijiazhuang 050003, China
    3.Military Representation Department, Xianyang, Shaanxi 712099, China
  • Online:2017-04-01 Published:2017-04-01


李  胜1,张培林1,李  兵2,吴定海1,胡  浩3   

  1. 1.军械工程学院 七系,石家庄 050003
    2.军械工程学院 四系,石家庄 050003
    3.总装西安军事代表局 咸阳地区军事代表室,陕西 咸阳 712099

Abstract: To fasten the speed of parameters updating and simplified steps of genetic operation in quantum genetic algorithm, a novel algorithm of Quantum Genetic Algorithm with Universal Quantum Gates(UQGA) is proposed. In this method, firstly, universal quantum gate is the unit of logical computation for chromosomes. Then, Hadamard gate is used for basic operation. Each gene is selected and mutated with universal quantum gates through a novel function of rotation angle. Finally, by the solution of fitness function, the global optimal set is got. Meanwhile, the algorithm is mathematically proven to be convergent. The algorithm is applied in function extremum search and feature selection for Iris dataset. The experimental results indicate that, compared with CQGA and CGA, UQGA has better performance in global search and feature selection, especially in convergence speed, execution time and classification accuracy.

Key words: quantum computation, universal quantum gates, Quantum Genetic Algorithm(QGA), function extremum search, feature selection

摘要: 为加快量子遗传算法的参数更新速度,简化遗传操作步骤,提出了一种基于通用量子门的量子遗传算法(Quantum Genetic Algorithm with Universal Quantum Gate,UQGA)。该方法以通用量子门为逻辑计算单位,对染色体进行遗传操作。利用Hadamard门进行基础变换;通用量子门通过新的旋转角度函数,对各个基因位进行选择、变异操作;通过求解适应度函数,得到全局最优解;同时,算法经数学证明是收敛的。该算法应用到函数极值搜索和Iris数据集特征选择中。实验结果表明,UQGA具有较好的全局搜索和特征选择性能,尤其是在收敛速度、运算时间和分类准确率方面明显优于普通量子遗传算法和普通遗传算法。

关键词: 量子计算, 通用量子门, 量子遗传算法, 函数极值搜索, 特征选择