A usable selection range standard based on test suite reduction algorithms

Abstract

In this paper, we analyze the features and distinctions of 6 classical algorithms: greedy algorithm (G), greedy evolution algorithm (GE), heuristics algorithm (H), greedy heuristic G (GRE), integer linear programming algorithm (ILP) and genetic algorithm (GA) to ensure the main influencing factors—the performance of algorithms and the running time of algorithms. What’s more, we would not only present a research design that aims at gaining deeper understanding about the algorithm classification and its function as well as their distinction, but also make an empirical study in order to obtain a practical range standard that can guide the selection of reduction algorithms. When the size of a test object (product of test requirements and test cases) is smaller than 2000×2000, G algorithm is the commonly recommended algorithm. With the growth of test size, the usage of GE and GRE becomes more general.