元胞贝叶斯决策路径规划方法  

Path planning of cellular Bayesian decision

作　　者：刘佶鑫[1] 赵英凯[1] 

机构地区：[1]南京工业大学自动化学院,江苏南京210009

出　　处：《计算机工程与设计》2009年第17期4053-4056,共4页Computer Engineering and Design

基　　金：国家863高技术研究发展计划基金项目(2006AA040308-02)

摘　　要：提出了基于贝叶斯决策的元胞自动机局部路径规划方法。以元胞蚁群系统能够求取离散化地图的全局最优路径,但是该方法缺乏对局部环境变化的适应性。将元胞蚁群系统得出的状态先验概率数据,以贝叶斯决策理论求取最小条件风险用于最优路径识别,在此基础上运用提出一个新概念--超级风险,可以处理一类环境改变后的局部路径规划。实验结果显示该方法可行且具有一定的智能化。Cellular automata path planning based on Bayesian decision theory is presented. The best path of discrete map is gained by using cellular ant colony system. But the disadvantage of this method is devoid of adaptability of environment. The state transcendental data of the CACS can be used to identified the best path by adopting Bayesian decision theory. And a new conception is created to resolve a kind of partial path planning in variational environment. The result of experiment shows this new method is feasible and intelligent.

关 键 词：元胞自动机 蚁群系统 贝叶斯决策 离散优化 路径规划 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]