Self-Organizing Fuzzy Control Based on Modified Artificial Fish-Swarm Algorithm

This paper presents modified artificial fish swarm algorithm (MAFSA) for automated design and optimization of self-organizing fuzzy logic controller (SOFLC). MAFSA has main improvement in the information of global best AF which is added to the behaviours of AF. It is proposed a novel method of adaptive step and visual based on food concentration. These improvements not only increase the capability of global searching, but also have considerable advantages like high convergence speed, flexibility and high accuracy. Then on the basis of MAFSA, SOFLC is designed and optimized by seven parameters, including three correlation factors of membership functions, one modifying factor of fuzzy logic rules, two quantitative factors and one scaling factor. The fitness function as optimized objective is defined by comprehensive performance index of the controller. Finally, simulation in temperature control of dyeing machine is implemented by using the above proposed method. Moreover, the simulation results show that MAFSA based SOFLC can avoid premature effectively and possesses strong adaptability and high control precision.