A Test Generation Method for EFSM-based Protocols Using the Transitions Feasibility Estimation

Abstract

Although extensive studies have been done on the protocol conformance testing based on an extended finite state machine (EFSM) model, the automatic generation of feasible test sequences is still a challenging task. A generated test sequence may be infeasible, due to the conflicts among the transition conditions and context variables of EFSMs. This paper proposed a test sequence generation approach for EFSM-based protocols conformance test by using the transition feasibility estimation. Firstly, our method generates candidate transition paths from a directed graph G which is derived from the EFSM model under testing ignoring all the predicates; Then, we designed a fitness function to guide the test generation with a trade-off among path feasibility, coverage criterion and path length. Finally, an adaptive exploration algorithm is developed to generate executable test sequences through expanding CPs. An experiment was designed to validate the effectiveness of the proposed method with two classic EFSM models. The experimental results show that our approach is more effective by comparing it to the TEA method based on breadth-first search (BFS).