改进的模拟退火算法在多目标跟踪中的应用研究  

Improved Simulated Annealing Algorithm for Multi-target Tracking

作　　者：蒋峥峥[1] 王丹丹[1,2] 陈晓红[1] 李一名 

机构地区：[1]南通大学计算机科学与技术学院,南通226000 [2]南通市紫琅职业技术学院,南通226002 [3]上海宝钢集团,上海200122

出　　处：《计算机与数字工程》2013年第11期1741-1743,共3页Computer & Digital Engineering

基　　金：南通市应用研究计划项目(编号:BK2012035);南通大学自然科学基金项目(编号:12Z036)资助

摘　　要：将模拟退火算法用于多目标跟踪中的数据关联分析,得到全局最优解。通过最近邻选择策略缩小模拟退火迭代过程中产生的解空间,提高算法效率。给出了改进之后的基于最近邻选择策略的模拟退火关联算法全过程并证明了该算法必包含最优解。对比仿真实验表明了改进后的算法在运算速度和关联性能上的优越性。Simulated annealing algorithm used in the data association analyzing of multi-target tracking can get the best solution. Nea- rest neighbor tactic was used to reduce the solution space generated by iterations of simulated annealing process, which improved the efficien- cy of the algorithm. An improved simulated annealing algorithm based on nearest neighbor tactic was proposed. The whole process of the al- gorithm was given and it was proved that the algorithm contains optimal solution. The simulation results revealed the superiority of the im- proved algorithm in the operation speed and associated performance.

关 键 词：多目标跟踪 模拟退火 数据关联 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]