论文部分内容阅读
将Petri网理论引入到神经网络模型的研究中，依据生物学神经系统的可塑性，通过适当扩展规范Petri网，建立了人工神经元状态转化的Petri网模型．在此基础上给出了利用遗传算法设计变结构神经Petri网的方法，该方法不仅可以得到满足要求的网络参数，而且能对网络的拓扑结构进行优化．仿真结果验证了该方法的可行性．
Petri nets theory is introduced into the research of neural network model. Based on the plasticity of biological nervous system, a Petri net model of artificial neuron state transition is established by properly extending the canonical Petri net. On this basis, a method of designing variable structure neuro-Petri nets by using genetic algorithm is given. This method can not only meet the requirements of network parameters, but also optimize the network topology. Simulation results verify the feasibility of this method.