Applying a MAX-MIN Ant System with a Dynamic Roulette Wheel Strategy to Software Release Planning
Qing Huang Yu, Wen Chiang Chuan, Hsu Huang Cheng · 2013
In this study, the software releasing planning (SRP) problem resulted from incremental software development is considered.Software releasing planning has been proven to be a NPcomplete problem.Owing to the intractable nature of the problem, a heuristic approach based on ant colony optimization (ACO) can be applied to obtain satisfactory suboptimal solutions within a reasonable amount of computational cost.However, most realistic ACO-based approaches for software releasing planning still remain to be improved.A novel ant-inspired search algorithm is therefore proposed.The proposed algorithm, namely MMAS DRW -SRP, adopts a dynamic roulette wheel strategy for giving a sophisticated balance between intensification and diversification, thereby improving the quality of solutions obtained.The performance of MMAS DRW -SRP is demonstrated by comparing it against conventional ACO-based approaches.Experimental results indicate that the proposed MMAS DRW -SRP algorithm performs significantly better than the competitive approaches.