Study On Quantum Genetic Algorithm Based On Mixed Update Strategy

Quantum genetic algorithm is a newly developed probabilistic evolutionary algorithm. Ithas the merit of relatively small population, powerful global research ability and rapidconvergence. However, when it solves complex optimization problems, its capacity isrelatively weak.In order to improve the performance of the algorithm for solving continuous functionoptimization problem, mixed update strategy and quantum genetic algorithm using a mixedupdate strategy are first proposed in this paper. The algorithm is inspired from mixed strategyand integrates different evolution operators into the quantum genetic algorithm. To verify theperformance of the algorithm for solving continuous function optimization problem, we selectdifferent representative functions in the standard test functions set to test the new algorithm.To compensate for the shortcomings of the capacity for solving complex optimizationproblems, a heterogeneous coevolutionary quantum genetic algorithm is proposed in thispaper which introduces the estimation of distribution algorithm from the perspective ofcoevolution. The population is divided into two sub-populations. One applies quantum geneticalgorithm using a mixed update strategy, and the other applies estimation of distributionalgorithm. Because it uses inter-generational migration operation as two sub-populationsinteraction, it not only retains the advantages of quantum genetic algorithm using a mixedupdate strategy, but also takes advantages of the estimation of distribution algorithm to solvecomplex combinatorial optimization problem.Finally, the heterogeneous coevolutionary quantum genetic algorithm is used to solve theknapsack problem. At present, many problems of industry and financial investment fields canbe transferred into the knapsack problem. Multiple groups of experiments show the efficiencyof the heterogeneous coevolutionary quantum genetic algorithm in solving knapsack problem.