Using design of experiments with discrete event simulation in operational research: a review

Design of experiments (DoE) is a branch of applied statistical methods, sometimes referred to as experimental design, that aims to efficiently assess the impact of input parameters on a system's responses. DoE covers a range of techniques that can help modellers and end-users better understand the system under different scenarios and support the identification of parameter configurations that lead to superior performance. While potentially useful for a better understanding of the interplay between different parameters in discrete-event simulation (DES) models, DoE is rarely used in such contexts. We conducted a review of the literature and identified 86 papers that used DoE techniques to analyse DES models. We found that the majority of these articles applied simple DoE techniques, namely full-factorial designs, to case studies. However, more advanced methods have also been combined with DES, both in methodological and applied articles. We present a case study that illustrates how DoE can be incorporated with a DES model in a healthcare context and enhance our operational decision-making. We review the benefits of using different experimental designs and metamodels in the context of a patient flow model through an enhanced care unit.