Improve The Verification Method Of Intelligent Optimization Algorithm Performance Test And Its Necessity

As a frontier science,the intelligent optimization algorithm solves the problem by solving the target problem and comparing the obtained results.At present,it has an unreasonable status quo: the algorithm with higher performance is continuously proposed,but the actual In the application,only classical optimization algorithms are widely used,and those with better performance are more in the paper.In fact,there are some shortcomings in the way of testing the performance of the algorithm,which needs further supplementation.Therefore,this paper proposes a verification method for improving the test performance and explains its necessity.First,introduce the methods commonly used to test the performance of the algorithm.Selecting nine classic standard test functions as targets,centered on the Gray Wolf Optimization Algorithm(GWO),and Cuckoo Search(CS),Chicken Swarm Optimization(CSO),Differential Algorithm(DE),Gravitational Search Algorithm(GSA)comparison.In the same environment,through experimental analysis,it is concluded that GWO has the best performance and far exceeds the rest of the algorithms.Then,it is pointed out that the algorithm that fails the verification tends to a certain preset point,and is no longer the optimal value of the target problem.This is missing from the method of testing performance and will mislead the direction of the algorithm.Therefore,this paper proposes one.The verification method was used to improve it in the same experimental environment,to verify GWO,and concluded that it did not pass the verification.After that,the goal of the verification method is to test whether the intelligent optimization algorithm is sensitive to the optimal solution location of the problem,and to explain that this method is a prerequisite for determining whether the method of testing the performance of the algorithm is effective.Afterwards,two kinds of improvement schemes were proposed for the cumbersome problem of this verification method,and it was tested that CSO and many algorithms outside this paper were not verified.Finally,by analyzing the basic principles and operation methods of GWO,PSO and CSO algorithms,it is determined that different intelligent optimization algorithms have different frameworks,which may cause different problems.Some even cannot determine the cause of the problem,and conclude that there is no unified correction.The most simple and effective way to solve the problem of intelligent optimization algorithm is to propose the verification method of this paper.