Abstract: Structural test is one of the basic approaches for test case identification. Because of the complexity of programming languages and the variety of applications, an efficient approach for automatic generation of structural test case is by iteration which is guided by program execution results. To examine the feasibility of full path coverage test, a R_N(K) path coverage criterion is presented. First, the test is treated as multiobjective optimization. To enhance the efficiency improvements are made on the encoding, the fittness function and the annealing gene of the genetic annealing algorithm. Then the optimization results and instumentation theory are used to generate test data. Experimental results show that this approach is better than the random method and the genetic method.

Key words: computer software, structural test, genetic simulated annealing algorithm, test case, control flow path table