Test Generation Based On EFSM Amorphous Slicing

Model based testing is a significance dimension in software testing. Extended Finite State Machine (EFSM) is extended from Finite State Machine (FSM). It has been noticed a crucial research significance and practical utility in EFSM based testing.EFSM based test generation includes test paths generation and test data generation that covers the test paths. Relevant scholars have studied on factors that affecting the efficiency of test generation in EFSM model, where the results show that the bigger the model scale, the lower the efficiency of test generation. Model based slicing is original from program slicing, it is a reduction of model. Among them, amorphous slicing is a kind of slicing technique with a high degree of reduction.In order to improve the efficiency of test generation in EFSM, amorphous slicing based test generation is proposed. According to characteristic of amorphous slicing and characteristic of path feasibility in EFSM, we present slicing based test path generation rules, and use genetic algorithm to generate test data. In addition, based on the definition of slicing, test case generated from slice is a subset of test case frome the original model. To ensure the test case generated on slice can be used in original model, we further propose the test path compensating rules and test data compensating rules. Two cases studies are introduced to show that test generation based on model slicing is more accurate in feasible path generation and test intensity improvement. In order to further verify the validity and correctness of the proposed technique, the experimental design and verification of the method are conduted.The results show that all of the test case generated from slice can be used in the original model, and in most cases, test generation efficiency on slice is higher than that on the original model. The main reason for low efficiency is that the time of dependence analysis and slicing is longer, but the overall evaluation time is very short.