Software Project Risks Management: Applying Extended Fuzzy Cognitive Maps with Reinforcement Learning

Abstract

During the last decade, risk management of Software projects اhas become a priority. Risk Management is defined as a whole process for identifying, analyzing and controlling risks in projects or organizations. It is used in all stages of project planning and control purposes during execution phase. The principal objective of research in this field is to be able to resolve conflicting problems through preventive intervention. We focus here on the adaptation of the organization to its environment by endowing it with the means that help’s to face to all unexpected in an intelligent way. Fuzzy Cognitive Maps emerged as a powerful tool for studying dynamic interactions in complex systems. There are two manners to construct them, the first one by experts of domain and the second by learning methods. In this paper, we develop a learning extended fuzzy cognitive maps based on a reinforcement learning algorithm so called Q-learning and we give an improve formulation of Kosko causality principle. This connection allows us to choose, based on learning historical data process, the best and the most important connections between concepts. In this work, we illustrate the effectiveness of the proposed approach by modeling and studying project risk management as an economic decision support system.