计算机工程与应用 ›› 2016, Vol. 52 ›› Issue (1): 151-156.

• 模式识别与人工智能 • 上一篇    下一篇

基于PCNN的迷宫路径搜索的评估函数优化研究

林  俊,刘  龙,谷  兵   

  1. 湖南师范大学 工程与设计学院,长沙 410081
  • 出版日期:2016-01-01 发布日期:2015-12-30

Research on evaluation function of path optimization search using PCNN

LIN Jun, LIU Long, GU Bing   

  1. College of Engineering and Design, Hunan Normal University, Changsha 410081, China
  • Online:2016-01-01 Published:2015-12-30

摘要: 在前人对PCNN模型的研究及应用的基础上,结合启发式的搜索策略——A*搜索策略,设计了基于改进型的PCNN迷宫智能优化算法,并将其应用解决实际迷宫问题。主要工作为:(1)通过对PCNN模型内在机理的研究,并根据PCNN的自身特点,选择合适的模型参数以适合求解迷宫最短路径问题。(2)选择与改进了的PCNN模型相结合的A*搜索算法,并证明该算法是可靠的,具有一定的自适应能力和所求得的解是最优解。(3)通过IEEE标准迷宫和MATLAB平台,对该算法的评估函数进行设计、仿真和验证等,不仅论证了(2)的结论,同时也论证了该算法的高效性。相关研究工作的实验结果表明,该算法不仅可以尽快找到目标,而且可以在相对少的搜索区域里得到相对满意的路径。

关键词: 脉冲耦合神经网络(PCNN), 迷宫路径, 优化算法, MATLAB平台, 评估函数

Abstract: On the basis of previous research and application of PCNN model heuristic search strategy—A*search strategy, designs the intelligent optimization algorithm on the basis of improved PCNN and its application to solve the practical problems in maze. The main tasks:(1)Through the research of the intrinsic mechanism of PCNN model, select the appropriate model parameters to fit the solving maze shortest path problem. (2)Combining A*search algorithm, and prove that the algorithm is reliable, adaptive capacity and the obtained solution is optimal. (3)Through the IEEE standard maze and MATLAB platform, the algorithm evaluation function of its design, simulation and verification, etc. not only demonstrate the conclusion(2), as well as demonstrate the efficiency of the algorithm. The related experimental results of research work show that the algorithm can not only find the target as quickly as possible, but also can get relatively satisfactory path in the relatively few search area.

Key words: Pulse Coupled Neural Network(PCNN), labyrinth path, optimization algorithm, MATLAB platform, evaluation function