An Optimal Optimization of Software Development Cost Estimation Using Genetic Algorithm

Estimating effort allows managers and software engineers to accurately anticipate, forecast, and quote schedule, budget, and manpower requirements. Determining estimated project cost, duration, and maintenance efforts well in advance Of the development stages is the biggest defiance to be attained for software projects. Formal models for cost estimation, such as the Constructive Cost Model (COCOMO) are limited by their inability to manage uncertainty in software projects early development cycle. This research presents an optimal optimization of software development cost estimation using genetic algorithm. It provides a fine solution to set the uncertain and ambiguous properties of software factors. It adopts COCOMO II model formulas and manages fine-tuning parameters for accurate effort and scheduled time software cost estimation. An experiment has been carried out using a NASA data set in order to improve the proposed algorithm. The experimental results show a significant optimization for both software effort up to 97.27% accuracy and scheduled time up to 98.88%. This research has emphasized that applying the genetic algorithm with fine-tuning parameters of COCOMO II model definitely improve the software development cost estimation.