Shuffled Frog Leaping Algorithm Improvement and Simulation Research for Optimization of Control Parameters
|School||Harbin Institute of Technology|
|Course||Control Science and Engineering|
|Keywords||Shuffled Frog Leaping Algorithm Quantum leapfrog algorithm PID control Fuzzy Control Parameter optimization|
The 2000 mixed leapfrog algorithm is a new kind of intelligent optimization algorithm, the basic idea of ??it comes from the the cultural genes heritage, its distinguishing feature is the hybrid cooperative search strategies with local search and global information. After a large number of simulation tests show that the shuffled frog leaping algorithm has superiority to solve high-dimensional, morbid, multiple local minima and other functions, is an effective optimization techniques. Choose from the basic principles of mixed leapfrog algorithm, steps and parameters of algorithm, the simulation test through multiple unconstrained benchmark functions with the typical characteristics of mixed leapfrog algorithm with particle swarm optimization, genetic algorithm two classical algorithm comparison, to examine the optimization of the performance of the algorithm. Mixed leapfrog algorithm has the advantage of easy to understand, and fewer parameters, but there is not easy to escape from local optima, slow convergence problem. For these shortcomings, this paper put forward the idea of ??three improved algorithms: the first two algorithm-based local search strategy improved algorithm design. A population the best individual and local best individual both to a certain proportion of random to affect cultural genome worst frog individual, the establishment of a new individual evolutionary formula, to update the worst individual position. Another for each independent evolutionary local search and update several poor individual fitness, so the cultural genome as a whole to quickly reach the optimal position. Two algorithms for the local search part of a single evolution given the improvements, simulation results show that two improved strategies to improve the solution speed of the algorithm, the validity has also been strengthened. Then quantum algorithm with mixed leapfrog algorithm combining quantum leapfrog algorithm. Probability amplitude of quantum bit structure frog individual quantum revolving door to update the individual to change the phase of the qubit quantum NOT gate population the best individual variation, and ultimately optimizing. Tested its fast convergence, solving a high success rate, effectiveness. In order to verify the quantum leapfrog algorithm and basic mixed leapfrog algorithm to optimize performance, control engineering field and use them to solve the PID controller parameter tuning and fuzzy controller parameters optimization problem. By comparing the simulation results with particle swarm optimization, genetic algorithms, quantum leapfrog algorithm has a satisfactory performance in the control field.