The Research Of Test-Suite Reduction Technology Based On Ant Colony Algorithm

Software testing is one of the very important parts in the process of software development.With the growing of software,the time,energy and materials spent on software testing are also increasing. The number of test suite and test execution costs of test case determines the cost and efficiency of software testing.In the assurance of quality and on the premise of fully tested the key operating of the software,how to use a minimum execution cost and number of test suite to test software fully then to lower the cost of software testing and to raise test efficiency,is the key point in this paper.There are two ways to resolve the problem of test-suite reduction:the one is the selection of test case;the other is the technique of test-suite reduction.The selection of test case is selecting a subset from the original test-suite,which could cover all repaired function.But the subset could not provide the same test coverage as the original test-suite.This paper will discuss the second one,which could fund minimum execution costs subset of the original test-suite.The minimum execution costs subset has the same testing coverage as the original test-suite.In order to minimize the cost of software testing,when reducing the test-suite,we must consider the number of reduction test cases and the execution costs of test cases.It is noted that the execution cost of each test case is not equal.There are many solutions to this problem:greedy algorithm,HGS algorithm,GRE algorithm,and so on. However,these methods relatively have a number of defects.For example,the solution accuracy of the greedy algorithm and HGS algorithm are low,and the solution speed of the GRE algorithm is slow.Based on summarizing and analyzing the existing algorithms for the test requirements,we provide an algorithm of test suite reduction which can get similarly the smallest price of the test suite. By using mutation index of genetic algorithm,this method based on the basic ant colony algorithm increases the searching randomness,rapidness and global convergence.It also overcomes the defects of precocity and stagnation.Implementing the algorithm is quick and simple,and the algorithm could find the balance point between the speed and the accuracy of solution.At last,we carry out a large number of experimental simulations to determine the parameter values of ant colony algorithm in the test suite reduction problem.After comparing the others classical algorithms,the experimental results show that the algorithm provided in this paper combining genetic algorithm with ant colony algorithm could get the test suite quickly which is similarly the smallest execution price and could be effective in reducing the cost of software testing.