【作者】 余迎； 聂仁灿； 周冬明； 赵东风；

【Author】 She Ying Nie Rencan Zhou Dongming Zhao Dongfeng (Department of Communication Engineering,Information College,Yunnan University,Kuraning 650091,China)

【机构】 云南大学信息学院通信工程系；

【摘要】 在竞争型脉冲耦合神经网络(CPCNN,competitive pulse-coupled neural metwork)的神经元一定参数设置条件下,利用其脉冲波高速并行传播可求解网络的最短路径。将迷宫等效为一固定连接权值的网络,利用 CPCNN 网络成功求解了迷宫的最短路径问题。实验仿真表明,CPCNN 网络求解迷宫最短路径所需的迭代次数只和入口到出口的最短路径有关,而和迷宫的大小和复杂程度无关,它利用较少的计算迭代次数即可得到迷宫的最短路径,具有较好的适用性。更多还原

【Abstract】 Under definite parameters setting condition of nerve cell for competitive pulse--coupled neural network(CPCNN)which features high speed parallel running,the shortest path of the network can be found.To equal the labyrinth into a network with fixed linking weight,the problem of the shortest path of the labyrinth is solved successfully using CPCNN.The simulation result shows that CPCNN needing itera- tion times for finding the shortest path in labyrinth is dependent of the length of the shortest path between the entry and exit in the labyrinth,and independent of the size and complexity of the labyrinth.Our algo- rithm can find the shortest path of maze with the minimum effort and has preferable practicability.更多还原