Research And Applications Of Particle Swarm Optimization Algorithm Based On Membrane Computing

Membrane computing is a newly distribution and parallel computing model, all classes of those computing devices considered in membrane computing are now generically called P systems. The research area of membrane computing originated as an attempt to formulate a model of computation motivated by the structure and functioning of a living cell, which is with the characteristics of parallel, nondeterministic and distributed. Most of the initial research problems and results were related to the computing power of various types of P systems, and most classes of P systems turned out to be equivalent to Turing machines, hence computationally complete/universal. Research on applications of P systems started relatively late, but it is now a very active research direction. Application areas include biology and bio-medicine as well as optimization, economics, and computer science. Particle swarm optimization algorithm(PSO) is an optimization algorithm which simulates the movement and flocking of birds. The theory of particle swarm optimization is incomplete, and there exist deficiencies on premature convergence and difficulty of application in discrete problem. So, the research on the theory analysis, improvement and discrete problem of particle swarm optimization is very significant.Considering the high parallelism and non-determinism of membrane computing, we introduce it as a basic computing frame which is do good to implementing particle swarm and discrete particle swarm optimization. In this paper, the first chapter of all is introduction. This chapter contains the research source and current situation of the membrane computing and particle swarm optimization. The second chapter of the paper is the recommendation of the fundamental theory, mainly demonstrating the principal theory and method of the membrane computing, PSO as well as discrete PSO.In third section, there are two main points contained in the work. Firstly, we proposed the dynamic selected-curve inertia weight particle swarm algorithm. In the DSCIWPSO, we set up several inertia weight curves, and used the rules in membrane computing to find out which curve is the most suitable one. Newly in the membrane model is the using of active membranes in our computing model, which is used to get two sub-swarms merged.So far in chapter four, we have introduced discrete particle swarm optimization. We got to know its properties first and sorry to find out that it had the deficiencies on premature convergence. We used a new idea to improving the probability mapping function, and at the same time change the position renew method on the promise of convergence.The last part of the paper is the summary and outlook, including the entire contents of the paper, at the same time, the shortcoming of this paper as well as the further resolving problems of P system based particle swarm structure are shown.