摘要
提出了一种改进的量子—粒子群算法来改善维数束缚问题。对于存在高维问题的量子—粒子群算法,引入了相互学习方法,使用多个粒子群用来优化解向量的分量,从而帮助粒子群克服维数束缚找到最优解;另外在每一次迭代过程中根据遗传算法中适应度函数对参与相互学习的粒子解的数目进行最优选取,从而有效减少了时间花费。对经典函数的测试计算表明,改进的混合算法确保了搜索精度,在时间花费上也得到了较好的改善。
An advanced quantum-particle algorithm was presented to solve the problem of the curse of dimensionality. The quantum-particle algorithm which has high-dimensional problem used multiple swarms to optimize different components of the solution vector by co-learning method and helped the algorithm break away from the curse of dimensionality to find the global optimal solution. Furthermore, in every iteration, we adopted the fitness function from genetic algorithm to select the number of the particles to reduce the time. The experimental results of classic functions show that the improved hybrid method keeps the balance between the global search and local search.
出处
《计算机应用》
CSCD
北大核心
2007年第12期2885-2887,共3页
journal of Computer Applications
基金
国家自然科学基金资助项目(60474030)
关键词
粒子群
量子
相互学习
适应度函数
particle swarm
quantum
co-learning
fitness function
