An Improved Approach for Class Test Ordering Optimization using Genetic Algorithms

Identifying the order in which the application classes have to be tested during the integration testing of object-oriented software systems is essential for reducing the testing effort. The Class Integration Test Order (CITO) problem refers to determining the test class order that minimizes stub creation cost, and subsequently testing effort. The goal of this paper is to propose an efficient approach for class integration test order optimization using a genetic algorithm with stochastic acceptance. The main goal of the class integration test order problem is to minimize the stubbing effort needed during the class-based integration testing. In our proposal, the complexity of creating a stub is estimated by assigning weights to different types of dependencies in the software system's Object Relation Diagram. The experimental evaluation is performed on two synthetic examples and five software systems often used in the literature for the class integration test ordering. The results obtained using our approach are better than the results of the existing related work which provide experimental results on the case studies considered in this paper.