Tutorial on optimisation via simulation: How to choose the best set up for a system

Currie, Christine and Monks, Thomas (2021) Tutorial on optimisation via simulation: How to choose the best set up for a system. Fakhimi, Masoud, Boness, Tom and Robertson, Duncan (eds.) In Operational Research Society 10th Simulation Workshop, SW 2021 - Proceedings. Operational Research Society. pp. 35-41 . (doi:10.36819/SW21.004).

Record type: Conference or Workshop Item (Paper)

Abstract

In this tutorial we consider the problem of finding the best set up to use for a system, where the objective is measured using the output of a stochastic simulation model. What makes this a difficult problem is that the output is stochastic and consequently changes in each replication. Optimisation via simulation is a vast topic and we restrict ourselves to a small part of it-ranking and selection-in which a small number of discrete options are being compared. We describe two of the best-used methods, KN++ and OCBA. In these algorithms, just one solution is returned at the end of the optimisation and there is a single objective. We also discuss variations including best subset selection, multi-objective optimisation via simulation, and the minimisation of the expected opportunity cost. The tutorial is accompanied by a Github repository which includes Python code for the algorithms we describe here.

This record has no associated files available for download.