Research On Path-Oriented Constraint Solving For Logical Expression

The development of path-oriented test case generation solves the problems of manual test in software testing to a certain extent such as low efficiency and high risk.It has become a hot research area in software testing.The establishment and solving of constraint system is an important part in path-oriented test case generation,and the method of solving constraints efficiently is a difficult problem in automatic generation of test cases.Firstly,according to the different coverage criteria,the framework uses different algorithms:in the criteria of Block Coverage and Branch Coverage,it uses the improved hill-climbing method based on the binary search tree that built as the initial hill-climbing space.In the criteria of Modified Condition/Decision Coverage(MC/DC),it transformed the logical expression constraint into atomic expression constraint set depend on independent influence pairs to solve it.The obtained variables and the reduced domain mapped to variables are used as input of the test case generation in branch and bound framework.Secondly,some optimization strategies are proposed.Using truth table optimization strategy and heuristic contradiction handling strategy to improve the ability of constraint solving in the criterion of MC/DC;using pruning algorithm to reduce the scale of binary search tree;using the sorting algorithm of atomic expression constraints based on variable dependence to enhance the ability of domain reduction;using infeasible path detection techniques to reduce the cost of unnecessary solving processes.The author has implemented these methods in Code Test System(CTS).The experiments show that the proposed strategy can support the constraint solving of logical expression in complex project and it also can successfully improve the code coverage in unit test.