A Learning Multiple-valued Logic Network Using Genetic Algorithm

This paper describes a genetic algorithm based learning Multiple-Value Logic (MVL) network. The proposed learning network operates on a population of candidate window parameters to produce new window parameters with lower errors between the desired outputs and the actual outputs of the MVL network. Thus, the learning MVL network has a large number of search points, making it possible to obtain a global minimum. The learning capability of the proposed MVL network with genetic algorithm is confirmed by simulations on several typical MVL functions. The simulation results show that the genetic algorithm based learning MVL network efficiently finds the appropriate network, window parameters, and bias, so that the MVL functions, especially for those relatively small problems.