Decision Maker Preferences in Multi-objective Stochastic Simulation Optimization: applications in operations management.
This Ph.D. dissertation focuses on the field of stochastic multi-objective optimization. Previous research in our group has shown the power of Gaussian Processes (GPs) for multi-objective optimization in settings with stochastic outcomes (such as, e.g., objectives measured through stochastic simulation). This Ph.D. research will focus on including the decision maker's preferences in the search for Pareto-optimal solutions, as well as in the identification of the points among which (s)he is indifferent. As such, this PhD dissertation is on the interface of optimization and machine learning; the hypothesis is that the use of GPs can be helpful to focus the search on the area of the Pareto front that is most relevant for the user, by saving (expensive) computer simulation time.
Period of project
01 December 2019 - 30 November 2023