摘要
一般邻域搜索方法面临着邻域定义的难点:定义的邻域较小,搜索就可能很快陷入局部最优,相反,则搜索效率会显著下降.针对这一问题,提出了一种基于极坐标的快速邻域搜索算法.试验证明,该算法能有效地解决邻域定义问题,并能在一定程度上解决常用的优化方法还较难解决的非凸集问题,对于一般复杂度问题,有较小的时间和空间复杂度.
Definition of adjacent domain is a difficulty that the arithmetic for adjacent domain search is facing. The small adjacent domain leads to local optimization, while the contrast leads to low efficiency. A polar coordinates-based agile arithmetic is proposed in this paper for adjacent domain search to solve the above difficulty. The tests demonstrate that the novel arithmetic can define the adjacent domain well,and solve the concavity problem in a certain extent, which is still difficult for normal optimization methods. The novel arithmetic has a lower complexity of time and space for a general problem.
出处
《武汉理工大学学报（交通科学与工程版）》
2006年第4期660-663,共4页
Journal of Wuhan University of Technology（Transportation Science & Engineering）
基金
湖北省自然科学基金青年杰出人才基金项目资助(批准号:2005ABB023)
武汉市晨光计划项目资助(批准号:20055003059-4)
关键词
极坐标
邻域搜索
快速算法
polar coordinates
adjacent domain search
agile arithmetic
作者简介
罗亚波：男，33岁，博士后，副教授，主要研究领域为优化设计、虚拟设计。
