A Novel Framework for Test Domain Reduction using Extended Finite State Machine

Abstract

Test case generation is an expensive, tedious, and error- prone process in software testing. In this paper, test case generation is accomplished using an Extended Finite State Machine (EFSM). The proper domain representative along the specified path is selected based on fundamental calculus approximation. The pre/post-conditions of class behavior is derived from a continuous or piece-wise continuous function whose values are chosen from partitioned subdomains. Subsequent test data for the designated class can be generated from the selected test frames. In so doing, the domain is partitioned wherein reduced test cases are generated, yet insuring complete test coverage of the designated test plan. The proposed modeling technique will be conducive toward a new realm of test domain analysis. Its validity can also be procedurally proved by straightforward mathematical principles.