Anti-Predatory NIA Based Approach for Optimizing Basic COCOMO Model

Abstract

Software Effort Estimation (SEE) is an important activity during development and production of software projects. The estimated effort is directly associated with the various planning and financial activities. It is also directly associated with business success. Constructive Cost Model (COCOMO) is a widely accepted SEE model. But in the current development scenario, existing parameters of COCOMO don't give realistic results. In the recent past, many researchers improved the performance of COCOMO by optimizing the parameters with the help of various Nature-Inspired Algorithms (NIAs). In this paper, a recently proposed NIA which is based on the frog's anti-predator behavior is used for the optimizing the parameters of basic COCOMO for SEE of 18 software projects listed in NASA data set. The performance of the Anti-Predatory NIA (APNIA) based proposed approach is also evaluated on NASA18 software data set in terms of the Mean Absolute Error (MAE). The result obtained shows 93.41% improvement in terms of MAE as compared to the basic COCOMO, 40.69% improvement as compared to Genetic Algorithm (GA) and 0.93% improvement as compared to Particle Swarm optimization (PSO) with inertia weight in effort estimation by proposed approach.