Multi-objective ranking and selection: Optimal sampling laws and tractable approximations via SCORE. Issue 1 (2nd January 2020)

• Record Type: Journal Article
• Authors: Applegate, Eric A.; Feldman, Guy; Hunter, Susan R.; Pasupathy, Raghu
• Is Part Of: Journal of simulation. Volume 14:Issue 1(2020)
• Page Start: 21
• Page End: 40
• Publication Date: 2020-01-02
• Subjects: Multi-objective -- ranking and selection -- simulation optimization; Operations research -- Periodicals; Mathematical models -- Periodicals; Simulation methods -- Periodicals; 511.805
• DOI: 10.1080/17477778.2019.1633891
• Languages: English
• ISSNs: 1747-7778

Abstract

ABSTRACT: Consider the multi-objective ranking and selection (MORS) problem in which we select the Pareto-optimal set from a finite set of systems evaluated on three or more stochastic objectives. Toward determining how to allocate a simulation budget among the systems, we characterise the asymptotically optimal sample allocation that maximises the misclassification-probability decay rate, and we provide an implementable allocation called MO-SCORE. The MO-SCORE allocation simultaneously controls the probabilities of misclassification by exclusion and inclusion, identifies phantom Pareto systems crucial for computational efficiency, and models dependence between the objectives. The MO-SCORE allocation is fast and accurate for problems with three objectives or a small number of systems. For problems with four or more objectives and a large number of systems, we propose independent MO-SCORE (iMO-SCORE). Our numerical experience is extensive and promising: MO-SCORE and iMO-SCORE can be used to solve MORS problems involving several thousand systems in three and four objectives.