摘要
为提高组搜索优化算法求解多维函数优化问题的性能,提出一种结合逐维搜索、Metropolis准则、反方向视角和禁忌表策略的改进组搜索优化算法.逐维搜索策略逐维更新并评价成员位置,在每一维,更新的值与其他维组成候选位置,使用模拟退火的Metropolis准则来决定是否接受候选位置.反方向视角策略使成员按一定的概率做反方向搜索,禁忌表策略避免生产者始终保持不变.这些策略能更好地平衡算法的集中性和多样性.在典型测试函数上进行了仿真,结果表明改进策略是有效的,提高了组搜索算法求解多维函数优化问题的全局寻优能力和收敛速度.
To improve the performance of group search optimizer for multi-dimensional function optimization problem,this paper presents an improved version which is based on dimension by dimension search,Metropolis rule,reverse direction angle,and tabu list strategies.The dimension by dimension search strategy updates and evaluates member's position dimension by dimension.On each dimension,a candidate position is constructed by the updated value and the values of other dimensions,and then the Metropolis rule of Simulated Annealing algorithm is used to decide whether to accept the candidate solution.The reverse direction angle strategy allows members to search new positions in reverse direction angle with some probability.The tabu list strategy is used to prevent the producer from being the same always.Using those strategies,the proposed algorithm can get better balance between intensification and diversification.The simulation experiments,which were carried on benchmark functions,show that those strategies are effective,and they improve the global optimization ability and convergence speed of group search optimizer for multi-dimensional function optimization problem.
出处
《小型微型计算机系统》
CSCD
北大核心
2013年第3期611-616,共6页
Journal of Chinese Computer Systems
基金
福建省自然科学基金项目(2009J05043
2011J05044
2008J0316)资助
