A neural network approach to circuit topology generator

In this paper, we present a new approach to automatically generate circuit topologies based on a neural network (NN). The main advantages of this approach is that the system can easily generate the circuit topology with less designer´s experience within a reasonable time by using less computer resources. The main kernel of neural network part consists of two parts, the topology imagination engine and the topology filtering engine. The topology imagination generates circuit topologies based on the NN concept. On the other hand, the topology filtering engine is used for filters unsimulatable circuits and unsuitable circuits. In the last stage, the optimizer is employed in order to obtain the final solution of the circuit´s parameters of the results which come from our topology generator. An electronic analog circuit design example is shown