Identification of an irrigation station using hybrid fuzzy clustering algorithms based on particle swarm optimization

Abstract

Fuzzy c-means (FCM) and Gustafson-Kessel (GK) algorithms are the best popular fuzzy clustering techniques in terms of efficient, straightforward, and easy to implement. However, these algrithms are sensitive to initialization and easy to trap in the local minimum. The important issue is how to avoid getting a bad local minimum value to improve the cluster accuracy. In fact, the particle swarm algorithm is strong global searching ability which is based on swarm operation, it doesn't easily get into the local minimum and has a fast convergence speed. In order to overcome the weakness of traditional clustering algorithms and takes advantage of PSO, we integrate FCM and GK algorithms with fuzzy particle swarm algorithm (FCM-PSO and GK-PSO algorithms). In this paper, hybrid fuzzy clustering algorithms based on FCM, GK and PSO called FCM-PSO and GK-PSO are presented. A comparative study between the clustering algorithms is investigated to identify the parameter of irrigation station. Experimental results applied to the irrigation station show that the GK-PSO algorithm is more effective and robust compared to the other algorithms.