基于蚁群神经网络的模拟电路故障诊断研究

On Analog Circuit Fault Diagnosis Based on the Ant Colony Neural Network

为了克服BP算法收敛速度慢、易陷入局部极小点的不足,提出将蚁群算法用于模拟电路故障诊断的神经网络模型学习算法,通过对实际模拟电路的仿真测试,表明该模型能有效地提高包括容差在内的多故障的模拟电路的故障诊断准确率和诊断速度,取得了令人满意的应用效果。

This paper sets forth the neural network model learning algorithm in which ant a!gorithm is used for analog circuit fault diagnosis to overcome the deficiency of BP algorithm convergence speed＇s being slow and tendency to be involved in a minimal point. In addition, the simulation test of actual analog circuit has indicated that the model can efficiently improve the speed and the accuracy rate of diagnosing the failure of many analog circuit faults, including the circuit tolerance, and has achieved satisfactory effects.