Artificial Intelligence methods played a vital role in the research of software engineering areas. In particular, Software Testing, a part of Software Development Life Cycle (SDLC) plays a crucial role in quality software development uses the application of artificial intelligence techniques. We need a new testing approach that has both human like intelligence and at the same time completely automated. This is achieved by applying Intelligent Agents in the software testing activity and that in turn helps in identifying only the optimal few test cases which will reduce the total time and cost needed in the testing process. We applied blackboard based learning for the agents to act and react in the specified environment. In our proposed approach, we developed a framework called Intelligent Test Suite Optimizer which consists of two agents namely Intelligent Test Path Optimizer Agent (ITPOA) and Intelligent Test Case Optimizer
Agent (ITCOA) for test path and test case optimization respectively. The test adequacy criterions used are all state, branch and statement coverage criterion. Finally, we compared our results against existing algorithm Ant Colony Optimization (ACO) in test optimization and proved that our approach out performed ACO.
Keywords: Intelligent Agents, Software Testing, Test optimization, Software Quality, Software under Test (SUT)
[...] Replace a test case with the effective test case with highest mutation score by the following condition: If mut_score(testcasei) > mut_score(testcasej) then replace the testcasei with testcasej. Else retain the test case testcasei. Step Generate the next set of test cases based on the previously selected test cases. Step Repeat steps 1 to 4 till the termination condition is reached. The test cases that have the highest coverage criterion and mutation score are selected and stored in the repositories. V. [...]
[...] Proposed Approach The Intelligent Agent based test optimization framework consists of the following agents: Intelligent Test Path optimizer Agent (ITPOA) Intelligent Test Case optimizer Agent (ITCOA) The agents have several Knowledge Bases and a Rules engine which will be controlled by a controller in them Controller that mediates among these knowledge sources. As per Nii, purpose of the black board is to hold computational and solution state data needed by the knowledge sources. The black board consists of objects from the solution space. [...]
[...] References D.Jeya Mala, V.Mohan, “Intelligent Test Seqeunce Optimization using Graph Based Searching Technique”, Proc. ICISTM, pp.10- Briand, L. the many ways Software Engineering can benefit from Knowledge Engineering”, Proc. 14th SEKE, Italy, pp. Dorigo M., Maniezzo, V., Colorni, A., Ant System: Optimization by a Colony of Cooperating Agents”, IEEE Transactions on Systems, Man, and Cybernetics-Part Vol No.1, pp.29- Horgan, J., London, S., and Lyu, M., “Achieving Software Quality with Testing Coverage Measures”, IEEE Computer, Vol No.9 pp. 60- Kit, Edward, “Software testing in the real world improving the process”, AddisonWesley Howe, A. [...]
[...] Design of Intelligent Test Path Optimizer using Blackboard Architecture The problem here is to generate the optimal test paths from the collection of infinitely many test paths in the SUT. At the high level, we have the objects such as Black board, Knowledge Sources and Controller. At the next level, the agent identifies the domain specific classes and objects that specialize these higher level abstractions. Test Paths Knowledge source 1 Knowledge source 2 Knowledge source ‘n' Controller States / Nodes Statements in the states Fig Blackboard architecture for Intelligent Test Path Optimizer Agent A. [...]
[...] SUT ITPOA ITCOA Level 1 Test Path Optimization Knowledge Source 1 Level 2 Test Case Optimization based on Paths Test path Optimization Test Case Optimization Knowledge Source 2 Level 3 Complete test Optimization Test Path Repository Test Case Repository Knowledge Source n Controller Fig Overall Black Board Architecture for the proposed framework Test Suite Repository Fig.3 Intelligent Test Suite Optimization Framework using Intelligent Agents A. Black Board Architecture In our proposed approach, we applied black board architecture for learning in Artificial Intelligence. [...]
using our reader.