Hybrid Pso Algorithm In The Optimization Of Nonlinear Control Systems And Circuit Parameters

Parameter optimization is one of the most important research objectsin many scientific, engineering problems and social economic activities.Currently, scholars have proposed a large number of optimizationalgorithms and Particle Swarm Optimization (PSO) is a novel, advancedevolutionary optimization algorithm among them. PSO algorithm hasbeen applied to many engineering problems successfully and achievesgood results.Auto-adaptive inertia weight PSO with mutation (PSOGA)algorithm and MPSODE algorithm with nonlinear inertia weight whichwas combined by differential evolutionary algorithm (DE) and PSO wereproposed in this paper. Those proposed algorithms aim to increasediversity of particle swarm and avoid evolving into local optimal point.PSOGA, MPSODE and PSO were encoded in real, which can avoidpotential problems, such as initial value sensitive and non-derivative,when using traditional algorithms to solve nonlinear problems. MPSODE,PSOGA and PSO algorithms were used on parameter optimization ofsemiconductor, typical instances in circuit and control system in order tocertify feasibility, effectiveness and advantage of the novel idea andproposed hybrid optimization algorithm in application.Linear mathematic models of semiconductors, circuits and controlsystem's nonlinear parts were constructed in this paper at first. Thensuitable parameters of mathematic models for optimization wereextracted and felicitous fitness functions of those models were established,which reflected the capability of parameter combination accurately.Under reasonable design flow, an efficient parallel search in theparameters target space were implemented to obtain the optimalparameter by utilizing the excellent search performance and fastconvergence speed of PSO and hybrid optimization algorithms proposed.Simulation experiments were used to certify that, compared withorther algorithms, the proposed hybrid algorithms can receive moresatisfied results of semiconductor capability, circuits consumption,decade in stopband of Infinite Inpulse Respose (IIR) filter, step response of PID control system and so on when those algorithms were used inparameter optimization process of semiconductor, nonlinear circuitswhich contain semiconductor, and PID controller. Comparisions of thosedifferent optimization algorithms in convergence speed, precision, andstability were also presented.