Particle Swarm Optimization Membrane Algorithm And Its Application Research

As a new branch of natural computing, membrane computing, inspired by biological cells, is a computing model which employs various features specific to the structure and functionality of the living cells. Membrane computing is parallel, nondeterministic and distributed. Theoretically, most membrane computing systems are computationally universal theoretically. Thus, it is very important to further explore the theory of membrane computing. However, compared with the theoretical study, membrane computing is a rather new research direction with a well-defined practical interest, and therefore further studies are very necessary to extend the use of P systems for real-world applications. Consequently, membrane algorithms are researched and applied in bench function optimization, broadcasting problem in P systems and time-frequency atom decomposition of radar emitter signals to expand the application of membrane computing. It is significance in theory and application. The main work and research fruit s are as follows.1. A particle swarm optimization based on P systems (PSOPS) is introduced, and is used for bench function optimization, broadcasting problem in P systems and time-frequency atom decomposition to promote the application of membrane computing. PSOPS combines the framework and evolution rules of P systems with PSO. In the elementary membrane, PSO is employed to evolve the system. The communication rules are used to exchange the information among individuals. Experiments carried out on bench function optimization, broadcasting problem in P systems and time-frequency atom decomposition of radar emitter signals show that PSOPS has greater global search ability and faster convergence speed than the counterpart PSO. It can improve the efficiency and achieve higher success rate in processing broadcasting problem in P systems. At the same time, it can effectively extract the best atom features from an over-complete time-frequency atom dictionary to decrease the computational complexity.2. To improve the optimization capability of PSOPS, by introducing the wavelet mutation into PSOPS to search the best solution, HPSOPS, a combination of P systems approaches and hybrid particle swarm optimization with wavelet mutation (HPSOWM), is proposed. Experiments carried out on bench function optimization, broadcasting problem in P systems and time-frequency atom decomposition of radar emitter signals show that HPSOPS can obtains significantly better solutions and has faster convergence rate than PSOPS,PSO,HPSOWM and GA. It can improve the efficiency more effectively and achieve higher success rate in processing broadcasting problem in P systems. At the same time, it can more effectively extract the best atom features from an over-complete time-frequency atom dictionary to decrease the computational complexity.This work was supported by the National Natural Science Foundation of China (60702026) and the Scientific and Technological Funds for Young Scientists of Sichuan (09ZQ026-040).