| Optimization problems widely exist in the fields of engineering and science.Its goal is to select appropriate variables under relevant constraints to optimize the objective function.Compared with the traditional optimization algorithms,which need function derivation and the solution results depend on the initial value,the meta heuristic algorithm proposed in recent years has less restrictions on the objective problem.The objective function can be differentiable or non differentiable,continuous or discontinuous,constrained or unconstrained,convex or nonconvex.Imperialist competitive algorithm(ICA)is an swarm intelligent optimization algorithm inspired by the social and political phenomenon of imperialist competition.It shows good optimization performance in photovoltaic parameter estimation.However,the algorithm sometimes falls into local optimization,and there are still some problems,such as insufficient balance between exploration and exploitation.Therefore,this dissertation studies the imperialist competitive algorithm as follows:Firstly,aiming at the problem that the assimilation process of imperialist competitive algorithm can not better explore the optimal solution,combined with the idea of differential evolution,based on the evolutionary strategy with adaptive differential mutation(ADM),an improved algorithm ICA-ADM of ICA is proposed.The strategy uses ”De/best/1” mutation operator and crossover operator,and adaptively calculates the scaling factor of mutation operator and the crossover rate of crossover operator according to individual fitness to generate new colonies.Because this strategy has stronger ability to explore the optimal solution,ICAADM has better exploration and exploitation ability,and can always maintain the diversity of solutions in the search process.The experimental results of 16 benchmark functions show that ICA-ADM can effectively solve the optimization problem,and the optimization performance is greatly improved.Then,ICA is extended to solve multi-objective problems.Based on this method,a multi-objective imperialist competitive algorithm(MICA)is proposed.In this method,an external archive is introduced to ICA to store the Pareto solution set.In order to enhance the exploration ability of the solution,the differential evolution strategy is used to search in the decision space,and the solutions in the population and archive are selected and updated through ensemble fitness ranking and non-dominated sorting,which can improve the diversity of the solution and the convergence of the algorithm.In order to verify the performance of the algorithm,the algorithm is tested on unconstrained function set(UF).Compared with other multi-objective algorithms,the results show that MICA can find the optimal Pareto front of multi-objective problem and has good results in solving complex multi-objective problem.In order to further verify the performance of the two improved ICA algorithms,they are applied to the photovoltaic parameter estimation problem and environmental/economic dispatch problem respectively.By minimizing the root mean square error between the PV cell model current and the calculated current,the estimation problem is transformed into an optimization problem,and the ICA-ADM is used to solve the problem to identify the parameters.The experimental results show that ICA-ADM can identify the parameters more accurately and have higher parameter stability,while MICA can successfully solve the environmental/economic dispatch problem. |
