Automated Test Data Generation For Programs Having Array Of Variable Length And Loops With Variable Number Of Iteration

摘要

In this paper, we propose a heuristic called longest path criterion for test data generation of programs having arrays of variable length, loops with variable number of iteration. Our heuristic is computed by a mathematical relation of array size(arr) and number of iteration(k). The relation we found is based on our experiment result at a saturation point after which no more longer path is executed. Our relation computed k can be found directly in stead of trial and error procedure for finding k. Test data with our k achieves better code coverage. We achieved 97% accuracy for predicting the value of k generating the longest path. The time required to generate test data for longest path criterion shows the effectiveness of our heuristic.