Improved Whale Algorithm And Its Application To Electric Vehicle Charging Optimization Problem

Due to the rapid growth of the engineering application and scientific computing scale,complicated systematic optimization problem has not satisfied solution is obtained by the traditional numerical method,in recent years,as the researchers for the further research of intelligent optimization algorithm,found that follow the animals and plants in the nature ecological habit of bionics algorithm of large-scale complex problem can be calculated precisely,It makes up for the deficiency of traditional optimization methods in solving large-scale complex problems and has a wider application field.Whale Optimization Algorithm(WOA)is a heuristic intelligent search algorithm proposed by Australian researcher Seyedali Mirjalili in 2016 inspired by the unique foraging behavior of humpback whales.The proposed algorithm has been successfully applied to site selection and path planning,cloud resource scheduling,photovoltaic power prediction,optimal power flow of power system,industrial design and other problems.WOA algorithm has some problems,such as unstable solution sometimes,easy to fall into local extremum and weak applicability to some engineering optimization problems.Aiming at the above problems,this paper improved and tested the whale algorithm on the basis of referring to a large number of materials and literature,and implemented the improvement and improvement of the algorithm,and applied the improved algorithm to the orderly charging of electric vehicles and the site planning of electric vehicle charging stations.The main work and innovations of this paper are as follows:(1)In order to solve the problems of whale optimization algorithm,such as poor stability,slow convergence speed and easy to fall into local extreme value,double population interactive evolutionary whale algorithm based on roulette and quadratic interpolation mechanism(DRQWOA)was proposed.The roulette selection mechanism is applied to the search and foraging stage to avoid the problem of multiple selection of inferior solutions.In the optimization process of evolutionary iteration,information exchange between two populations with different evolutionary strategies was carried out continuously.After the whale position changes,the second interpolation mechanism is used to update the whale position again,and the updated position is preferentially replaced.The algorithm flow is given in detail,the convergence of the algorithm is proved,and the DRQWOA time complexity analysis is given.Finally,a multi-dimensional comparison test was carried out on the CEC2017 test function set for six algorithms in the same software and hardware environment.According to the experimental data,the optimization performance of DRQWOA algorithm was significantly improved.(2)A whale optimization algorithm with quasi-opposition learning,real-time boundary processing and enhanced search mechanism(QREWOA)is proposed.In the improved algorithm,the judgment conditions of the random search strategy and the bubble net predation strategy are adjusted,and the optimal location information is introduced into the random search strategy.The quasi-opposition learning mechanism was introduced after the updating of individual position.The original boundary processing method is optimized to properly deal with a large number of homogeneity problems caused by boundary processing.Then,the fitness of QREWOA algorithm and the graduality of global optimal solution are proved theoretically,and the time complexity is analyzed theoretically.Finally,the CEC2017 test function set is compared with six representative algorithms.Experimental results show that QREWOA algorithm has higher solving accuracy and optimization stability than other five comparison algorithms on various complex functions.(3)The DRQWOA algorithm and QREWOA algorithm are applied to solve the orderly charging and charging station location problems of ev respectively.In view of the problems such as line overload,unstable operation of power supply network and increased peak-valley difference of distribution network load caused by electric vehicle disordered charging,DRQWOA algorithm is used to guide electric vehicle charging scientifically and rationally,and charging optimization scheduling is realized on corresponding target conditions.Six comparison algorithms are used to simulate the multi-objective mathematical model of ev charging optimization,and the results show that DRQWOA algorithm is more applicable and efficient in orderly charging.Aiming at the problem of planning layout and service scope division of urban electric vehicle charging stations,QREWOA algorithm and Voronoi diagram were used to solve the problem jointly.The simulation results of six comparison algorithms show that QREWOA algorithm can effectively solve the problems of uneven distribution of ev charging stations within the planning range and uncontrollable division of service range of charging stations.