A New Particle Swarm Optimisation Based Memetic Procedure for Fuzzy J-K Flop Neural Networks Learning

Fuzzy Flip-Flop Neural Networks are a type of neural network that combines the strengths of fuzzy logic and recurrent neural networks. In this article, a new method of training the Fuzzy Flip-Flop Neural Network is described. This combines classic particle swarm optimisation procedure with selected evolutionary procedures. Among these are mutation or crossing. The proposed algorithm is tested both on the example of regression of a simple function and then on benchmark data used for classification. In most test cases, the memetic procedure has been shown to provide fast convergence, high accuracy, and robustness, making it a popular choice for the learning and optimisation of Fuzzy Flip-Flop Neural Networks.