Providing a model to estimate the probability of the complexity of software projects

Abstract

Function Point Analysis (FPA) is most used technique for estimating the size of a computerized business information system which was developed by Allan Albrecht. Various studies proposed new methods to extent FPA algorithm; mainly they tried to make it more precise but they are based on the similarity of previous projects so this paper is proposed. This paper, presents a statistical simulation method that can be applied for each generic project. The proposed method is a new method to assess estimation of size and effort of software projects by a stochastic and Markov chain approach. Based on Metropolis-hasting simulation algorithm, we formulate a Probabilistic Function Point Analysis (PFPA). Moreover, A Bayesian belief network approach is used for determination of complexity of system. It determines the function weights utilizing Markov chain theory to support estimating the effort of software projects. As a case study, this new method is applied in online publication domain. This method can increase the chance of implementation of generic projects on time.