Research Of Quantum Genetic Algorithm Based On A Combination Of Multi-operator

Quantum genetic algorithm is a new optimization method,which has a very strong vitality and research value. Quantum algorithm incorporates many of the basic features of quantum mechanics, it has gradually become a new computing model and greatly improve the computational efficiency, Thus the combination of genetic algorithm also greatly improve the efficiency of genetic algorithm. But quantum genetic algorithm is easy to fall into local extreme point in complex continuous function optimization, it also has slow convergence and other shortcomings,some scholars have proposed a quantum particle swarm optimization algorithm to solve this problem, with good global convergence. But in solving the problem of premature convergence is still not very good, will produce suboptimal solution, affecting the performance of the algorithm. Therefore, the quantum particle swarm optimization based on multi-elitist strategy is proposed in this paper.First, in this paper, in the process of quantum particle swarm integration, we update the qubit state by particle swarm optimization, so updating the rotation angle will be updating the particle position.Secondly, on the basis of this, we also propose multi-elitist strategy, introduce the concept of growth rate and candidate area, which makes the particle trajectories is not guided by global optimal position, but depends on the promising candidate search area. The enhanced quantum particle swarm algorithm has better performance in some difficult optimization problems, and verified by experiment.Finally, the quantum particle swarm algorithm based on multi-elitist strategy is applied to the clustering problem. in cluster analysis, based on genetic algorithm, K-means algorithm both have the local search capability of K-means and global search capability of genetic algorithms, can better solve the problem of clustering. To further optimize the clustering effect, we introduce the multi-quantum particle swarm elite in K-means algorithm, which can minimize the differences within the cluster, but has a faster convergence rate, the algorithm combines the fast convergence capability of K-means and global search ability of multi-quantum particle swarm. In comparative tests with the K-means algorithm based on quantum genetic algorithm and so on, the K-means based on multi-quantum particle swarm elite has better performance.