多因子專案分群之最佳粒子群演算法實作軟體工作量預估

在資訊科技產業中，精準的評估每個軟體開發專案的工作量對於軟體公司在軟體開發成本與開發時程管理是非常重要的。從一個專案開始，多數的開發團隊都會覺得時間不夠用或是專案評估錯誤導致軟體專案失敗。然而軟體專案的成本幾乎都是人力成本，人力成本又與開發時程成正比，所以精準的工作量評估就更顯得重要了。因此，本研究將以皮爾遜相關積差係數分析( Pearson product-moment correlation coefficient ) 以及one-way ANOVA分析並選出數個因子經過K-Means分群演算法(K-Means clustering algorithm)對專案進行分群，再將分群後的專案利用N-1筆專案透過最佳粒子群演算法 (Particle Swarm Optimization , PSO)以平均誤差率(MMRE)做為適應值不斷的往最佳參數趨近，最後再利用最佳化後的參數引用至欲做預測的專案求出軟體工作量。本研究利用COCOMO 中的63筆歷史專案來進行測試，實驗結果確實表現出利用多因子做為專案分群依據能比COCOMO最初的三個模式更有效的預估軟體工作量。

關鍵字

軟體工作量；專案分群； K-Means分群演算法；最佳粒子群演算法；皮爾遜積差相關係數

並列摘要

In the IT industry, precisely evaluate the effort of each software development project to develop cost and development schedule management to the software company in the software are count for much. Since a project, majority of development teams will feel time isn't enough to use or the project valuation be false to make the software project failed. However the cost of the software project is almost a human resource cost, human resource cost and then become a direct proportion with development schedule, so precise effort the valuation more seem to be getting more important. Consequently, this research will use Pearson product-moment correlation coefficient and one-way analyze to select several factors then used K-Means clustering algorithm to software project clustering. After project clustering, we use Particle Swarm Optimization that take mean of MRE (MMRE) as a fitness value and N-1 test method to optimization of COCOMO parameters. Finally, take parameters that finsh the optimization to calculate the software project effort that is want to estimation. This research use 63 history software projects data of COCOMO to test. The experiment really expresses using base on project clustering with multiple factors can make more effective base on effort of the estimate software of COCOMO's three project mode.

並列關鍵字

software effort ； correlation coefficient ； Particle Swarm Optimization ； K-Means clustering algorithm ； project clustering

參考文獻

[Ho 2011] Wen-Hsien Ho, Agnes Lai-Fong Chan,” Hybrid Taguchi-Differential Evolution Algorithm for Parameter Estimation of Differential Equation Models with Application to HIV Dynamics” Mathematical Problems in Engineering, Vol. 2011, Hindawi Publishing Corporation, 2011, pp.???.

[Umapathy 2010] Prabha Umapathy, C. Venkataseshaiah, and M. Senthil Arumugam,” Particle Swarm Optimization with Various Inertia Weight Variants for Optimal Power Flow Solution”, Discrete Dynamics in Nature and Society, Hindawi Publishing Corporation, Vol.2010,2010, pp.???.

[Ahmed 2005] A. Moataz, Ahmed, Moshood Omolade Saliu , and Jarallah AlGhamdi,“Adaptive fuzzy logic-based framework for software development effort prediction”,Information and Software Technology Vol.47, Issue 1, 1 January 2005, pp.31-48.

[Chen 2003] Dechang Chen, Dong Hua, Jaques Reifman, and Xiuzhen Cheng, “Gene Selection for Multi-Class Prediction of Microarray Data” IEEE Proceedings of the Computational Systems Bioinformatics(CSB2003) ,2003, pp.???.

[Eveleens 2010] J.L. Eveleens,C. Verhoef, “The Rise and Fall of the Chaos Report Figures” Software, IEEE ,Vol.27, Issue 1, pp.30-36,2010.

國際替代計量

多因子專案分群之最佳粒子群演算法實作軟體工作量預估

未授權

主題瀏覽