空间数据关联的多目标粒子群优化算法  

Multiple objective particle swarm optimization algorithm with space data association

作　　者：王经卓[1] 樊纪山[1] 

机构地区：[1]淮海工学院电子信息工程学院,江苏连云港222005

出　　处：《控制与决策》2015年第7期1291-1297,共7页Control and Decision

基　　金：国家自然科学基金项目(61174013);淮海工学院自然科学基金项目(Z2012010)

摘　　要：提出一种空间联合概率数据关联的多目标粒子群优化(DS-MOPSO)算法.采用正态分布确保初始样本均匀分布,通过采用拥挤距离和先验概率采样确立外部归档中非支配解的拥挤度来保持解的多样性;采用Sigma方法作为选择精英粒子策略寻找全局最优解;利用空间联合概率数据关联动态生成每个粒子的惯性权值,增强粒子的搜索区域,防止算法陷入局部最优.仿真实验结果表明,采用所提出的算法所得到的Pareto解集具有很好的收敛性和多样性.A multi-objective particle swarm optimization algorithm based on data association （DS-MOPSO） is introduced. The Gauss distribution is adopted to ensure the uniform distribution of initial population. The crowding distance and the prior probability are used to calculate the crowing degree of non-dominated solutions in the external archive. The elite particles, which maintain the diversity of solution, are selected by the Sigma value. The space joint probabilistic data association is introduced to generate dynamically the inertia weight of each particle and expand the search area in order to avoid falling into local optimum. Simulation results show that the Pareto solutions obtained by that DS-MOPSO has a good convergence and diversity.

关 键 词：多目标优化 数据关联 粒子群优化 拥挤距离 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]