MO²TOS: Multi-Fidelity Optimization with Ordinal Transformation and Optimal Sampling

Abstract

Simulation optimization can be used to solve many complex optimization problems in automation applications such as job scheduling and inventory control. We propose a new framework to perform efficient simulation optimization when simulation models with different fidelity levels are available. The framework consists of two novel methodologies: ordinal transformation (OT) and optimal sampling (OS). The OT methodology uses the low-fidelity simulations to transform the original solution space into an ordinal space that encapsulates useful information from the low-fidelity model. The OS methodology efficiently uses high-fidelity simulations to sample the transformed space in search of the optimal solution. Through theoretical analysis and numerical experiments, we demonstrate the promising performance of the multi-fidelity optimization with ordinal transformation and optimal sampling (MO2TOS) framework.

A model-driven approach to interoperability between simulation and optimization for production and logistics systems

“…Multi-fidelity ordinal transformation methods use low-fidelity simulation models to evaluate more solutions, which then provide guidance for selecting solutions to evaluate using high-fidelity simulation models [46]. Multi-fidelity simulation methods for production and logistics systems have been studied, including when and how to to use system approximations [47][48][49][50].…”

A Stochastic Queueing Model for Capacity Allocation in the Hierarchical Healthcare Delivery System

“…Finally, beyond those articles that use analytical model, another stream of articles use simulation optimization model to addresses the complex referral process in the hierarchical healthcare system, such as Xu et al (2015Xu et al ( , 2016 and Song et al (2016). These articles do not address patients' decentralized choice behavior.…”

Accelerated Simulation of Discrete Event Dynamic Systems via a Multi-Fidelity Modeling Framework

“…In the second stage, low-fidelity analysis is replaced with a radial bias function approximation. Xu et al [33] presented an ordinal transformation approach to transform the original solution space into a one-dimensional space based on the rankings of solutions using the low-fidelity model. After that, they used an optimal sampling approach to search solutions in the transformed one-dimensional space for optimal solutions using high-fidelity simulations.…”