Research On Quantum Evolutionary Computational Methods And Its Application In Feature Selection

Evolutionary computational methods represented by swarm intelligence optimization algorithms have been applied to all aspects of life,such as solving shortest paths,Knapsack problems,and Traveling Salasman problems,et al.In recent years,many excellent swarm intelligence algorithms have been proposed,such as particle swarm algorithm(PSO),grasshopper optimization algorithm(GOA),et al.However,in solving real problems,these algorithms will expose various shortcomings.In order to solve the problems of dragonfly algorithm,mothflame optimization algorithm and slime mold algorithm,the thesis put forward improvement measures respectively according to the idea of quantum computation,and successfully applied to feature selection problem.The main research contents are as follows:(1)On the basis of dragonfly algorithm,quantum rotation gate and gaussian mutation mechanism are introduced to proposed dragonfly optimization algorithm(Dragonfly Algorithm Based on Quantum and Gaussian,QGDA).The improved dragonfly algorithm based on two mechanisms is put into the benchmark function for testing,and the results show that QGDA can obtain better solutions than other algorithms.In addition,QGDA can also achieve higher accuracy by conducting feature selection tests on different data sets.(2)In terms of moth-flame optimization algorithm,this thesis proposes an improved moth-flame optimization algorithm(Moth-Flame optimization algorithm based on Quantum and Simulated Annealing,QSFMO)based on quantum computation and simulated annealing mechanism.In the experiment,compared with the original MFO algorithm,the performance of QSMFO has been significantly improved.At the same time,QSMFO also shows a better effect in feature selection problem.(3)Finally,aiming at the shortcomings of slime mould optimization algorithm,an improved slime mould optimization algorithm(Slime Mould Algorithm based on Quantum and Water Cycle,WQSMA)based on quantum computation and water circulation mechanism is proposed.WQSMA is compared with 13 meta-heuristic algorithms in the CEC2014 test set,and the experimental results showed that WQSMA ranked first.In addition,the proposed algorithm can obtain higher classification accuracy after verification on multiple data sets in practical application.