摘要
为了进一步提高量子粒子群算法的精度,从描述粒子状态波函数的δ势阱特征长度L(t)出发,重新修改其评价方式。通过给群体中的每个粒子引入随机权重,生成随机权重平均最优位置来重新评价L(t),以增强算法的随机性,帮助算法逃离局部极小值点的束缚,使算法尽快找到全局极值点。通过几个典型函数测试表明,改进算法的收敛精度优于QPSO算法,并且具有很强的避免陷入局部极值点的能力。
In order to further improve the accuracy of Quantum Particle Swarm Optimization algorithm, the evaluation method of δ trap characteristic length L（t） of wave function for describing the particle’s state is modified. Introducing a random weight to each particle in swarm, and generating a random -weighed mean best position to reassess L（t） , enhance the algorithmic randomness, help algorithm to escape from local minima to manacle, make the algorithm to find the global extreme points. Through the test of several typical functions, its result shows that the convergence accuracy of the improved algorithm is better than QPSO algorithm’s, and it can be very strong to avoid falling into local extremums.
出处
《计算机工程与应用》
CSCD
2013年第10期25-27,共3页
Computer Engineering and Applications
基金
国家自然科学基金项目(No.50874094)
四川省教育厅重点项目(No.11ZA040)
西华师范大学博士启动基金项目(No.12B022)
关键词
粒子群优化
量子粒子群优化
随机权重
随机加权平均最优位置
Particle Swarm Optimization
Quantum-behaved Particle Swarm Optimization
random weight
random-weighted mean best position
