A Differential Evolution and Spatial Distribution based Local Search for TrainingFuzzy Wavelet Neural Network

Many parameter-tuning algorithms have been proposed for training Fuzzy Wavelet Neural Networks(FWNNs). Absence of appropriate structure, convergence to local optima and low speed in learningalgorithms are deficiencies of FWNNs in previous studies. In this paper, a Memetic Algorithm (MA) isintroduced to train FWNN for addressing aforementioned learning lacks. In proposed MA, DifferentialEvolution (DE) is utilized as the global search. The main contributions of this paper aresummarized inthree sections. (I) Proposing a new, fast and effective local search based on spatial distribution that isnamed Spatial Distribution Local Search (SDLS). SDLS can adjust the step size of movements towardbetter neighbor solutions adaptively. (II) Introducing a selection method to select appropriate individualsfrom current population for local refinement in MA. This property decreases the computational cost ofMA and leads to tuning the local search frequency in an adaptive way. (III) Improving the selectionoperator in standard DE by an adaptive strategy. In this strategy, worse offspring has a chance to bereplaced with its parent to prevent trapping in local optima and controlling the selection pressure. Theproposed MA is compared with several training algorithms of FWNNs over some benchmark problems.Experimental results obtained, confirm the effectiveness of the proposed MA for improving theconvergence rate and modeling accuracy in comparison to the other training methods