Generation of improved test cases from UML state diagram using genetic algorithm

Abstract

An UML statechart diagram is used to model the dynamic aspects of object oriented software systems. A Finite State Machine (FSM) and an Extended Finite State Machine (EFSM) are commonly used for Model Based-Testing (MBT). A test case generation from FSM is easy due to unconditional transition from state to state by traversal. A test case generation from EFSM needs to satisfy the guard condition before transition from state to state. Testing from EFSM is difficult due to unfeasible paths and test data. This paper proposes a search-based approach to find feasible transition sequences and test data generation. The Unified Modeling Language specifications are translated into extended flow graphs. An Extended flow graph from EFSM specifies the control and data flow in a statechart diagram. To guide the feasible transition path a genetic algorithm (GA) framework is proposed. It is shown that the test cases so generated yield an improved test set.