Research On Artificial Fish Swarm Optimization Algorithm Based On Membrane Computing

Membrane computing (P systems) is a novel natural computing method based on the membrane structure and the cooperation in membranes. According to the characteristics of the membrane structure, P systems can be divided into three types: cell-like P systems,tissue-like P systems and neural-like P systems. Membrane algorithm was first proposed by Nishida in 2004, when he introduced genetic algorithm into P systems. After that, membrane algorithm became one of the important research branches in membrane computing. An improved artificial fish swarm algorithm based on P systems (IAFSAPS) was proposed in this paper. Extensive experiments with benchmark functions and 0-1 knapsack problems were investigated by IAFSAPS. It was also applied to deal with the fault diagnosis problems of power systems. The main research contents are shown as follows:(1) IAFSAPS was proposed by introducing artificial fish swarm algorithm (AFSA) into one level membrane structure in P systems. And some communication rules were proposed according to the characteristics of the membrane structure and operators. The choices of some important parameters, number of basic membranes and the interval of information communication, were studied by solving some unimodal and multimodal functions. The experiment results in solving 29 benchmark functions and 7 0-1 knapsack problems with different number of items showed the searching ability, stability and the optimization speed of IAFSAPS.(2) The passive regions in power system was researched and the relationship between the tripped circuit breaker and the main protection, first backup protection and second backup protection was analyzed. From the perspective of intelligent optimization algorithm,one kind of strategy based on greedy heuristic algorithm was proposed. The passive regions could be found out according to the fitness. Then IAFSAPS was used to solve some single passive region and multiple passive regions problems, and all the passive regions were found out.(3) The fault diagnosis problem of power system was transferred into a kind of 0-1 programming problem. IAFSAPS was applied to solve this mathematical model. Many simple and complex fault problems were solved by IAFSAPS, the experiment results showed the stability and rapidity of IAFSAPS. Then by solving some typical fault diagnosis problems and comparing with the results in some references, IAFSAPS could accurately locate the fault positions in the power systems.