Class-integration Testing Order Research Based On Coupling Measures

Software testing is a software quality assurance mean with high intricacy and expensive cost.Owing to rapid development of object-oriented technology,the software testing faces with new challenges and difficulties.The main object-oriented technology features abstraction,inheritance,encapsulation and polymorphism makes the program more standardized structure,data security and improves code reuse rate to compare with traditional software.However,the object-oriented software testing difficulty increased due to the complex dependencies of objects.Studies have found that the subsystem level,namely class-integration testing is a difficult and focus in software testing.So how to find an optimal class integration testing order to reduce the cost of object-oriented software testing,thereby reducing the cost of enterprise has become an urgent problem.This paper firstly introduces the research background,domestic and overseas research status,the basics of object-oriented software testing as well as the similarities and differences with the traditional test,analyzes the detailed solution of integrated test issues raised by domestic and foreign scholars.Then,according to the pros and cons,proposes an approach based on coupling measures to assess the complexity of each stub,finds the object diagram all cycles,combined with the graph heuristic algorithm to ensure minimizing the overall complexity of test stub constructed by breaking cycles to obtain the optimal test sequence.Finally,it validates the feasibility of this method by examples,compared the result with famous scholars Briand et al.and Abdurazik et al.The experimental results show that this paper method has lower complexity and high efficiency under the same situation,lead to the lower testing cost and shorter test cycles of object-oriented software.