Date of Award


Document Type


Degree Name

Master of Science in Operations Research


Department of Operational Sciences

First Advisor

James W. Chrissis, PhD


Many problems exist where one desires to optimize systems with multiple, often competing, objectives. Further, these problems may not have a closed form representation, and may also have stochastic responses. Recently, a method expanded mixed variable generalized pattern search/ranking and selection (MVPS-RS) and Mesh Adaptive Direct Search (MADS) developed for single-objective, stochastic problems to the multi-objective case by using aspiration and reservation levels. However, the success of this method in approximating the true Pareto solution set can be dependent upon several factors. These factors include the experimental design and ranges of the aspiration and reservation levels, and the approximation quality of the nadir point. Additionally, a termination criterion for this method does not yet exist. In this thesis, these aspects are explored. Furthermore, there may be alternatives or additions to this method that can save both computational time and function evaluations. These include the use of surrogates as approximating functions and the expansion of proven singleobjective formulations. In this thesis, two new approaches are developed that make use of all of these previous existing methods in combination.

AFIT Designator


DTIC Accession Number