The Applied Research Based On Membrane Computing And Neural Network In MIMO System Control Field

The membrane computing is a branch of natural computing which abstracts computingmodels from the system and the function of living cells. It has the advantage of parallel anddistributed computing and great potential of research and development. The optimization andcontrol design based on the basic thought and research achievement of membrane computingand tree structure topological structure combined with neural network, which is applied toball-plate control system and so on. The simulations and comparison are of system are given infigure. The main contents of the thesis are as follows:(1)A tree-structure membrane optimization algorithm based on the basic thought andframework of standard membrane computing is researched in this thesis. It is divided into twoparallel subsystems branch. The rules of select, rewrite, transportation, crossover,communication are used.(2)Ball-plate system is a typically multivariable MIMO control system. In this paper, themembrane computing is applied to the optimization to the ball-plate system control systemand designed the fuzzy RBF neural network control scheme based on tree-structure membraneoptimization algorithm. The simulation of stabilizing control and trajectory tracking controldemonstrate the potential of the algorithm, compared with the control design optimized byPSO(particle swarm optimization) algorithm.(3)Then I also researched BP neural network PID control system based on tree-structuremembrane optimization algorithm of wind-tunnel. Compared the results with no optimized,got the conclusion and solve the problem.