PROMOTING ACCESS TO AFRICAN RESEARCH

Global Journal of Pure and Applied Sciences

Log in or Register to get access to full text downloads.

Remember me or Register



An improved COCOMO software cost estimation model

SOO Duke, JN Obidinnu

Abstract


In this paper, we discuss the methodologies adopted previously in software cost estimation using the COnstructive COst MOdels (COCOMOs). From our analysis, COCOMOs produce very high software development efforts, which eventually produce high software development costs. Consequently, we propose its extension, called Improved Benchmark for COCOMO software Cost Estimation (IBCOCOMO). Here, we implement the extension by
adopting development efforts in three perspectives . optimistic, pessimistic and most-likely. We carry out the implementation with four different software sizes . 10,000, 15,000, 20,000 and 50,000 lines of codes (LOC). When compared, the estimated values of the optimistic efforts (Eopt) estimated with COCOMO II using the sample sizes, IBCOCOMO model reduced it by 9.67%, 9.74%, 9.72% and 9.70% (. 10%) respectively. In order to ensure a realistic developmental effort favourable to both software developers and customers, a standard effort multiplication factor(er)
is introduced, to further calibrate and remove any unforeseen residual errors that may defect the effort due to human factors unseen or unsighted.



AJOL African Journals Online