Rapid goal-oriented automated software testing using MEA-graph planning

  • Authors:

  • Manish Gupta;Jicheng Fu;Farokh B. Bastani;Latifur R. Khan;I. -Ling Yen

  • Affiliations:

  • Department of Computer Science, University of Texas at Dallas, Dallas, USA 75083-0688;Department of Computer Science, University of Texas at Dallas, Dallas, USA 75083-0688;Department of Computer Science, University of Texas at Dallas, Dallas, USA 75083-0688;Department of Computer Science, University of Texas at Dallas, Dallas, USA 75083-0688;Department of Computer Science, University of Texas at Dallas, Dallas, USA 75083-0688

  • Venue:
  • Software Quality Control
  • Year:
  • 2007

Quantified Score

Hi-index 0.00

Visualization

Abstract

With the rapid growth in the development of sophisticated modern software applications, the complexity of the software development process has increased enormously, posing an urgent need for the automation of some of the more time-consuming aspects of the development process. One of the key stages in the software development process is system testing. In this paper, we evaluate the potential application of AI planning techniques in automated software testing. The key contributions of this paper include the following: (1) A formal model of software systems from the perspective of software testing that is applicable to important classes of systems and is amenable to automation using AI planning methods. (2) The design of a framework for an automated planning system (APS) for applying AI planning techniques for testing software systems. (3) Assessment of the test automation framework and a specific AI Planning algorithm, namely, MEA-Graphplan (Means-Ends Analysis Graphplan), algorithm to automatically generate test data. (4) A case study is presented to evaluate the proposed automated testing method and compare the performance of MEA-Graphplan with that of Graphplan. The empirical results show that for software testing, the MEA-Graphplan algorithm can perform computationally more efficiently and effectively than the basic Graph Planning algorithm.