Refrigerated Showcase Fault Detection by a Correntropy Based Artificial Neural Network Using Fast Brain Storm Optimization

Abstract

This paper proposes refrigerated showcase fault detection by a correntropy based Artificial Neural Network (ANN) using Fast Brain Storm Optimization (FBSO). Since there are approximately 50,000 convenience stores in Japan and it is difficult for experts to tune up all of showcase systems with different characteristics. Therefore, an automatic parameter tuning method for various showcase systems such as ANN should be applied. Effectiveness of the proposed method is verified by comparison with conventional least square error (LSE) based ANNs using stochastic gradient descent (SGD) and correntropy based ANNs using Differential Evolutionary Particle Swarm Optimization (DEEPSO) with actual showcase data.