一种渐变概率选择操作的实数编码量子进化算法  

Real-coded Quantum Evolutionary Algorithm Based on Transition Probability Selection Operation

作　　者：张伟丰[1] 钱洁[1] 

机构地区：[1]湖北汽车工业学院经济管理学院,湖北十堰442002

出　　处：《湖北汽车工业学院学报》2013年第4期53-57,共5页Journal of Hubei University Of Automotive Technology

摘　　要：为进一步提高实数编码量子进化算法在进化过程中的种群多样性以及在高维复杂函数优化上的全局收敛性,参照模拟退火算法的特点,提出了一种渐变选择概率的实数编码量子进化算法,该方法通过在进化过程中逐步提高更好解的选择概率,在进化计算初期保持种群的多样性,能较为全面地对解空间进行搜索,而在进化末期,选择概率逐渐提高到1,只接受更好的解而保证算法稳定的收敛。仿真实验结果表明,该算法能有效避免早熟和局部极值问题,具有更快的收敛速度和更高的求解精度。In order to further improve the population diversity in the process of evolutionary and convergence in the high-dimensional complex function optimization of real-coded quantum evolu- tionary algorithm, according to the characteristics of simulated annealing algorithm, a real-coded quantum evolutionary algorithm based on transition probability selection operation was proposed, in which through gradually raising the selective probability of better solutions in the process of evolutionary, the population diversity was maintained in the preliminary stage of evolutionary computation, the solution space could be searched more thoroughly, but late in evolution the selective probability was increased to 1 gradually, only the better solution would be accepted, thus the stability of convergence could be ensured. The simulation results show that this algorithm can avoid premature and local extreme, meanwhile, it has faster convergence speed and higher accuracy.

关 键 词：量子进化 实数编码 优化 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]