Synthesizing test scenarios in UML activity diagram using a bio-inspired approach

• Published in: Computer languages, systems & structures Vol. 50; pp. 1 - 19
• Main Authors: Arora, Vinay, Bhatia, Rajesh, Singh, Maninder
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.12.2017
• Subjects: Amoeboid organism; Bio-inspired algorithm; Computational intelligence; Concurrency; Test scenario generation; UML activity diagram; Amoeboid organism; Test scenario generation; UML activity diagram; Concurrency; Computational intelligence; Bio-inspired algorithm
• ISSN: 1477-8424; 1873-6866
• DOI: 10.1016/j.cl.2017.05.002

•We applied biologically inspired algorithm for test scenario generation in UML activity diagram.•Amoeboid organism algorithm is better as compare to its peer approaches like genetic algorithm, ant colony meta-heuristic.•For large search spaces, the amoeboid organism algorithm works better when compared with the exact algorithm like DFS.•Activity diagrams are taken from benchmark data set and student project having variety of control constructs. The model-based analysis is receiving a wide acceptance as compare to code-based analysis in the context of prioritizing and guiding the testing effort and speeding up the development process. Ordinarily, system analysts as well as developers follow Unified Modeling Language (UML) activity diagrams to render all realizable flows of controls commonly recognized as scenarios of use cases. This paper applies a bio-inspired algorithm to produce test scenarios for the concurrent section in UML activity diagram. Here, the heuristic draws its inspiration from the internal mechanism of the slime mould Physarum Polycephalum, a large single-celled amoeboid organism. Simulations are performed using eight subject systems taken from the LINDHOLMEN data-set, two models taken from real life student projects and five synthetic models. The results obtained through different approaches are validated through the statistical analysis which demonstrates that our proposed approach is better than the existing Ant Colony Optimization (ACO) and Genetic Algorithm (GA) by a number of feasible test scenarios generated.