A New Optimized Hybrid Model Based On COCOMO to Increase the Accuracy of Software Cost Estimation

The literature review shows software development projects often neither meet time deadlines, nor run within the allocated budgets. One common reason can be the inaccurate cost estimation process, although several approaches have been proposed in this field. Recent research studies suggest that in order to increase the accuracy of this process, estimation models have to be revised. The Constructive Cost Model (COCOMO) has often been referred as an efficient model for software cost estimation. The popularity of COCOMO is due to its flexibility; it can be used in different environments and it covers a variety of factors. In this paper, we aim to improve the accuracy of cost estimation process by enhancing COCOMO model. To this end, we analyze the cost drivers using meta-heuristic algorithms. In this method, the improvement of COCOMO is distinctly done by effective selection of coefficients and reconstruction of COCOMO. Three meta-heuristic optimization algorithms are applied synthetically to enhance the process of COCOMO model. Eventually, results of the proposed method are compared to COCOMO itself and other existing models. This comparison explicitly reveals the superiority of the proposed method.