Arasteh, Bahman2022-10-312022-10-312022Arasteh, B. (2022). Clustered design-model generation from a program source code using chaos-based metaheuristic algorithms. Neural Computing and Applications, doi:10.1007/s00521-022-07781-6https://doi.org/10.1007/s00521-022-07781-6https://hdl.handle.net/20.500.12713/3214Comprehension of the structure of software will facilitate maintaining the software more efficiently. Clustering software modules, as a reverse engineering technique, is assumed to be an effective technique in extracting comprehensible structural-models of software from the source code. Finding the best clustering model of a software system is regarded as a NP-complete problem. Minimizing the connections among the created clusters, maximizing the internal connections within the created clusters and maximizing the clustering quality are considered to be the most important objectives in software module clustering (SMC). Poor success rate, low stability and modularization quality are regarded as the major drawbacks of the previously proposed methods. In this paper, five different heuristic algorithms (Bat, Cuckoo, Teaching–Learning-Based, Black Widow and Grasshopper algorithms) are proposed for optimal clustering of software modules. Also, the effects of chaos theory in the performance of these algorithms in this problem have been experimentally investigated. The results of conducted experiments on the eight standard and real-world applications indicate that performance of the BWO, PSO, and TLB algorithms are higher than the other algorithms in SMC problem; also, the performance of these algorithm increased when their initial population were generated with logistic chaos method instead of random method. The average MQ of the generated clusters for the selected benchmark set by BWO, PSO and TLB are 3.155, 3.120 and 2.778, respectively.eninfo:eu-repo/semantics/closedAccessChaos TheoryHeuristic AlgorithmsModularization QualityModule Dependency GraphSoftware Modules ClusteringArticleWOS:0008653801000032-s2.0-85139475766Q210.1007/s00521-022-07781-6N/A