Software Process Simulation Modelling: Systematic Literature Review (2020)

Abstract

> Changes and continuous progress in logistics and productive systems make the realization of improvements in decision making necessary. Simulation is a good support tool for this type of decisions because it allows reproducing processes virtually to study their behavior, to analyze the impact of possible changes or to compare different design alternatives without the high cost of scale experiments. Although process simulation is usually focused on industrial processes, over the last two decades, new proposals have emerged to bring simulation techniques into software engineering. This paper describes a Systematic Literature Review (SLR) which returned 8,070 papers (published from 2013 to 2019) by a systematic search in 4 digital libraries. After conducting this SLR, 36 Software Process Simulation Modelling (SPSM) works were selected as primary studies and were documented following a specific characterization scheme. This scheme allows characterizing each proposal according to the paradigm used and its technology base as well as its future line of work. Our purpose is to identify trends and directions for future research on SPSM after identifying and studying which proposals in this topic have been defined and the relationships and dependencies between these proposals in the last five years. After finishing this review, it is possible to conclude that SPSM continues to be a topic that is very much addressed by the scientific community, but each contribution has been proposed with particular goals. This review also concludes that Agent-Based Simulation and System Dynamics paradigm is increasing and decreasing, respectively, its trend among SPSM proposals in the last five years. Regarding Discrete-Event Simulation paradigm, it seems that it is strengthening its position among research community in recent years to design new approaches.