量子演化算法的改进研究

An Improved Quantum-Inspired Evolutionary Algorithm

从两个方面对量子演化算法进行改进:(1)因量子染色体的量子位处于叠加态和纠缠态,无法使用传统的交叉方式对量子位进行操作,设计了针对测量后的经典染色体进行全干扰交叉,这样既不会破坏量子染色体的固有的并行性,又可以增加测量后染色体的多样性,继而影响量子染色体进化方向,加快算法的收敛速度,有效地防止"早熟";(2)设计了概率触发器启动量子非门进行量子变异。实验表明,改进的量子演化算法比起先前的算法具有更好的寻优能力,更稳定的收敛度。

In this paper,two aspects of the quantum-inspired evolutionary algorithm are improved：（1）Quantum chromosomes qubits in superposition and entangled states can not use the traditional cross-qubit operation,the full interference cross is designed for measured clas-sic chromosome,which could not undermine the inherent parallelism of quantum chromosome,and could increase the diversity of the chro-mosomes after the measurement,affect the direction of chromosome evolution,and effectively prevent ＂premature＂;（2）The probability trigger is designed to start the quantum NOT gate for quantum variation.The experiments show that the improved quantum evolution opera-tor has better search ability than the previous algorithm over into a more stable convergence degree.