A Path-oriented Automatic Random Testing Based on Double Constraint Propagation

A key issue in software testing is the actual generation of test data from program input domain. Obviously, more accurate input domain is, more efficient test generation is. This paper presents a path-oriented automatic random testing method based on double constraint propagation. For a given path, its domain can be reduced by splitting an input variable domain and executing a double constraint propagation algorithm. Moreover, a random test data generator is developed according to the reduced path domain and the test experiments are conducted on a number of programs. Experimental results show that the method gets more accurate path domain than PRT (path-oriented random testing) approach, and random testing efficiency can thus be enhanced by using the proposed method.