Journal of Optimization in Industrial Engineering Vol.13, Issue 2, Summer & Autumn 2020, 267- 278 883892.1700 10.22094/JOIE.2020.1 DOI: 267 Developing a Risk Management Model for Banking Software Development Projects Based on Fuzzy Inference System Tooraj Karimi a,* , Mohammad Reza Fathi a , Yalda Yahyazade a a Faculty of Management and accounting, University of Tehran college of Farabi, Iran, Tehran Received 30 November 2019; Revised 30 June 2020; Accepted 30 June 2020 Abstract Risk management is one of the most influential parts of project management that has a major impact on the success or failure of projects. Due to the increasing use of information technology (IT) systems in all fields and the high failure rate of IT projects in software development and production, it is essential to effectively manage these projects is essential. Therefore, this study is aimed to design a risk management model that seeks to manage the risk of software development projects based on the key criteria of project time, cost, quality and scope. This is presented after making an extensive review of the literature and asking questions from experts in the field. In this regard, after identifying the risks and defining them based on the dimensions and indicators of software development projects, 22 features were identified to evaluate banking software projects. The data were collected for three consecutive years in the country's largest software development eco-system. According to Rough modelling, the most important variables affecting the cost, time, quality and scope of projects were identified and the amount of risk that a project may have in each of these dimensions was shown. Since traditional scales cannot provide the accurate estimation of project risk assessment under uncertainty, the indexes were fuzzy. Finally, the fuzzy expert system was designed by MATLAB software that showed the total risk of each project. To create a graphical user interface, the MATLAB software GUIDE was used. The system can predict the risks of each project before each project begins and helps project managers be prepared to deal with these risks and consider ways to prevent the project from failing. The results showed that quality and time risks were more important than cost and scope risks and had a greater impact on total project deviation. Keywords: Project Risk Management; Software Development; Expert Systems; Rough Set Theory; Fuzzy Logi; Fuzzy Inference System. 1.Introduction In today’s ever-changing and complex world, projects and processes face immense risks that could cause major disruption in their progress, if not causing the entire project's failure, unless proper precautions and measures are taken into account in order to react to these risks. Risk management is a necessary measure for achieving goals in projects; hence, it is essential to pay special and precise attention to risk management (Akbarpour & Seyedesfahani 2010). Based on the project management body of knowledge, the risk is defined as any uncertain occurrence that would impact at least one of a project’s goals (PMI 2013). The software development process is prone to many risks, which is evident from the high rates of failure in such projects. Two of the major goals of a software development project is profit and meeting the deadlines. The presence of any risk in the project would result in additional costs and delay in its progress. However, a project’s success is not only related to time or cost, factors such as quality, performance, customers’ satisfaction and many others are also important indicators for a project’s success. A vast majority of previous models have been focused on costs and very few have considered quality and time, which results in the point that a small number of these models could apply to large projects (Zhang & Fan 2014). In this project, a model is designed that considers factors such as time, quality and scope of a project in addition to cost in risk assessments for software development projects. One of the challenges faced by the experts and supervisors of software development industry is lack of an intelligent predictive system for projects risk assessment. Among different systematic and analytic approaches, expert system (ES) has been acknowledged as an effective knowledge- based technique with a variety of applications in industry and services, e.g. failure prediction, performance evaluation and classification, disorder explanation, accident and fault diagnosis, process control and risk assessment (Ford, 1985). ES can play different roles, depending on the development extent of the knowledge base and the technology. Since ES is dependent on deduction, it must be possible to explain its reasoning to the solution in order to examine how it is argued. Given different uncertainties in software development projects, it is necessary to design a fuzzy model based on the fuzzy sets theory (Dokas et al., 2009). The fuzzy inference system provides a schematic process for converting a knowledge base into a nonlinear mapping. For this reason, knowledge-based systems (fuzzy systems) are used in engineering applications and decision making. The fuzzy inference system is a mapping of the input-output space that is implemented using membership functions and fuzzy rules (Nourian et al., 2019). Fuzzy sets and fuzzy logic are powerful mathematical tools for modeling uncertain industrial, human and natural systems. Fuzzy models facilitate decision making by means of approximate reasoning and linguistic terms. They play an important role when applied to the complex phenomena which are not easily *Corresponding author Email address: Tkarimi@ut.ac.ir