基于混沌映射的PSO及其在图像分割中的应用  

Chaos-map-based particle swarm optimization algorithm and its application in image segmentation

作　　者：田东平[1] 田絮资[1] 

机构地区：[1]宝鸡文理学院计算机科学系,陕西宝鸡721007

出　　处：《西北大学学报（自然科学版）》2012年第6期925-930,共6页Journal of Northwest University（Natural Science Edition）

基　　金：陕西省科技厅自然科学基础研究基金资助项目(SJ08-ZT13)

摘　　要：目的提出了一种基于混沌映射的粒子群优化算法。方法一方面,应用逻辑自映射函数初始化均匀分布的粒群以提高初始解的质量;另一方面,根据群体早熟收敛的判断机制,在算法进化过程中引入局部变异机制和局部重新初始化粒群的方法以有效避免算法陷入局部收敛的缺点。结果该算法应用在基准测试函数优化中能有效提高全局寻优的性能,且稳定性好;应用在图像分割中取得了与遗传算法同样好的分割效果。结论提出的算法具有有效性和实用性,可用于求解高维复杂函数以及工程优化问题。Aim A chaos-map-based particle swarm optimization algorithm （CMPSO） is proposed. Methods On the one hand, the uniform particles are produced by logical self-map function so as to improve the quality of the ini- tial solutions. On the other hand, according to the decision mechanism of the swarms＇ premature convergence, the local mutation mechanism and the local reinitializing particles are introduced during the evolutionary process in or- cler to help the PSO to break away from the local optimum. Results The CMPSO proposed in this paper is applied to two benchmark functions and the experimental results show that it can improve the performance of searching glob- al optimum efficiently and own higher stability. Meanwhile, the CMPSO is used in image segmentation and it also can get the good result as that the Genetic algorithm does. Conclusion The proposed CMPSO is effective and prac- tical. In addition, it can be applied to solve the high-dimensional complex functions and some engineering optimiza- tion problems.

关 键 词：粒子群优化 逻辑自映射 早熟收敛 变异机制 图像分割 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]