粗糙域A＊算法估价函数最优权研究  

Research on Evaluation Function Optimal Weight of A-star Algorithm Based Rough Area

作　　者：滑斌杰[1] 柴忠良[1] 林立忠[1] 

机构地区：[1]石家庄学院计算机系,石家庄050035

出　　处：《网络新媒体技术》2013年第2期12-17,共6页Network New Media Technology

基　　金：科技部科技型中小企业技术创新基金项目(编号:12(26211300753);河北省科技型中小企业技术创新基金项目(编号:11C130311104)

摘　　要：A*算法的核心问题是估价函数的构造及其权值确定问题。通常估价函数构造和权值确定主要依赖实验和经验法,这样构造的估价函数粗略,算法的复杂度较高且实时性差。为了解决这一问题,本文通过对粗糙域粗糙属性概率分布的分析,得出基于粗糙域A*算法估价函数的最优权与粗糙属性概率分布的标准差的相关性,并通过实验对结果进行了验证。实验结果表明最优权的确定使A*算法的复杂度明显降低,且能够满足实时应用的需要。The keys of A - star algorithm are build appropriate evaluation function and determine the weight of it. Usually, the building of valuation function and determining of weight depend on the experimental method, which makes the evaluation function rough, high complexity, and can not fit for real - time application. In order to solve this problem, this paper researchs deeply on probability distri- bution of rough area which A - star algorithm based and gets the correlation between roughness characteristics of rough area and the op- timal weight of evaluation function. Then, through many experiments the resuhs are verified. Experience result show that the optimal weight can achieve remarkable computation reduction and fit for real -time application.

关 键 词：最优路径 粗糙域 A~*算法 估价函数 标准差 最优权 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]