Research On Natural Calculation Method Based On Charge Characteristic Strategy And Its Application In Test Case Generation

With the rapid development of the information age,natural computing methods have been applied in many fields,and great breakthroughs have been made,including software testing,and how to efficiently select automated test case generation and effective combination of natural computing methods is very important.Aiming at the shortcomings of traditional natural calculation methods that are prone to fall into local optimum,poor algorithm performance,low solution accuracy and low test efficiency in software testing when solving multimodal and complex high-dimensional problems,this paper uses Coulomb’s law to propose a charge Kinematic characteristics strategy,and combined with natural computational methods to generate test case framework to solve the above problems.The main research results are:A natural calculation method of charge characteristic strategy is proposed,which is suitable for the natural calculation method based on population,and the execution process does not depend on the evolution process of the specific algorithm.This paper mainly uses the mutual movement of charged individuals in Coulomb’s law,and controls the movement of individuals to balance the exploration and development capabilities of natural computing methods,improve the chance of individuals jumping out of the local optimum,and effectively increase the diversity of the population.For the universality of the algorithm,the formula of Coulomb’s law and its parameters are designed.At the same time,the Markov chain is used to analyze the convergence of the charge characteristic strategy.The proposed strategy is applied to particle swarm optimization algorithm and differential evolution algorithm respectively,and its performance is verified by classical test function.At the same time,the fitness value and algorithm convergence rate are compared with several current mainstream algorithms.The experimental results show that the optimization accuracy and convergence rate of the natural calculation method after using this strategy are significantly improved.On the basis of natural computing method and software test case generation technology,a branch-coverage test case generation framework based on charge characteristics is proposed.Mainly using the effective combination of automatic generation of test cases and natural calculation methods,a new fitness function is designed.Through the analysis of branch distance function,branch nesting depth and branch weight,the new fitness function can be based on different branch conditions.With deep,adaptive computing fitness values,build a bridge between natural computing methods and test cases.In order to evaluate the effect of the generation framework of charge characteristics in the generation of structural test cases,it was applied to the particle swarm algorithm,and the benchmark program was used to conduct experiments to compare the particle swarm algorithm and the mainstream test case generation methods in recent years.The experimental results show that the charge characteristic framework has advantages in coverage and average evolutionary algebra compared with the unimproved original algorithm in the generation of test cases.