一类模糊神经网络的结构优化

提出了一类新的模糊神经网络结构，神经网络自构形学习算法用于此类模糊神经网络的规则推理层进行规则优化，并将此算法向网络的低层次上扩展，还可对此算法的各公式进行少许的改动和调整C1和C2的值相结合，达到网络结构优化的目的，最后以无人驾驶的模型小汽车为例进行软件仿真，结果分析，证明该方法是可行且有效的。

Structural Optimization of a Kind of Fuzzv-neural Network

The thesis constructs a kind of new fuzzy-neural network. Neural network self-structure learning algorithm is applied in the rule-reasoning layer of the fuzzy-neural network to optimize rules, and the algorithm is expanded in the layers under the rule-reasoning layer of the fuzzy-neural network, and modification of the formulas of the algorithm is combined with adjustment of C1 and C2 , so that the structure of the network can be optimized . Finally, the thesis executes software simulation and analyses the result by the example of no-person-driving small car model. The method proves to be effective and feasible.