COCOMO II-FG-HGWO: A Modified Hunting Gray Wolf Optimization for Improving Constructive Cost Model II Performance

Software effort estimation presents a significant challenge in the domain of project management. The Constructive Cost Model II (COCOMO II) is frequently applied to estimate the required software project effort. Our study proposed a Constructive Cost Model II with Fuzzy Gaussian and Hunting Grey Wolf Optimization (COCOMO II-FG-HGWO), which improves the accuracy of COCOMO II. Furthermore, the proposed method applies FG to obtain more optimal values of 11 COCOMO II effort multipliers. The GWO hunting mechanism is modified by adding tournament selection to obtain optimal alpha from populations by optimizing COCOMO II coefficients A and B. It helps each alpha explore their search abilities, thereby reducing the risk of being trapped in local optima. The experimental results show that the proposed method reduces MMRE by 0.01% on the NASA 60 dataset and achieves lower values by more than 16% compared to COCOMO II. This indicates that estimated effort is closer to actual effort, and the risk of error in calculating project costs becomes smaller and, in turn, improves the quality of software projects.