一种实数编码量子进化算法及其收敛性

Real-coded quantum-inspired evolutionary algorithm and its convergence

基于量子计算理论和进化理论,提出一种新的量子进化算法——基于实数编码的量子进化算法(RQEA).不同于传统进化算法的单点编码和量子进化算法的量子比特编码,该算法以实数矩形区域表示基因,一条染色体携带多个个体信息.利用量子态叠加和相干机理,通过叠加、变异及自学习来完成进化过程.理论分析证明了算法具有全局收敛性.实验结果表明,该算法在函数优化上具有优异的性能.

A novel algorithm, real-coded quantum-inspired evolutionary algorithm （RQEA）, is proposed based on evolution theory and quantum computation. Different from the representation of single point coding and quantum-bits coding, rectangle region is employed in RQEA to represent gene, and a chromosome fakes several individuals information. By using quantum computation mechanics, interference operator, mutation operator and self-learning operator are proposed to accelerate evolution process. Theoretical analysis shows that RQEA converges to the global optimum. Experiment results show that RQEA has a strong ability of global optimization and high convergence speed.