Analysis of Different Techniques for Optimizing COCOMOII Model Coefficients

Software effort estimation is an essential and important issue in the software industry. Earlier estimation is required by the project managers to allocate resources and time plan the project efficiently. Cost estimation of a project is usually based upon the person days, thus total cost can be calculated by multiplying daily person day rate with the number of persons employed on that project. So, the prediction of cost is totally dependent on effort estimation. Accurate estimation is very important in decision making process for software development. Various models have been proposed for effort estimation and COCOMOII model is one of them. It is a very popular and widely used method for estimating effort and time duration of the project. This model uses four coefficients namely a, b, c, d in its formulae with their predefined values. When these values of coefficients are put into the estimation formulas, they give good results but still far from the real values. To reduce the vagueness of results of estimation formulas, optimization of a, b, c, d coefficients is necessary. Thats why various techniques can be superimposed on COCOMOII model to improve its results such as simulation, neural network, genetic algorithm, soft computing, the fuzzy logic modeling, tabu search, simulated annealing etc. In this paper, we are going to study these techniques and try to find out which one is better.