Access Type

Open Access Dissertation

Date of Award

January 2021

Degree Type


Degree Name




First Advisor

Boris Mordukhovich


This thesis is devoted towards the study and solving of a new class of multi-facility location problems. This class is of a great theoretical interest both in variational analysis and optimization while being of high importance to a variety of practical applications. Optimization problems of this type cannot be reduced to convex programming like, the much more investigated facility location problems with only one center. In contrast, such classes of multi-facility location problems can be described by using DC (difference of convex) programming, which are significantly more involved from both theoretical and numerical viewpoints.In this thesis, we present a new approach to solve multi-facility location problems, which is based on mixed integer programming and algorithms for minimizing differences of convex (DC) functions. We then computationally implement the proposed algorithm on both artificial and real data sets and provide many numerical examples. Finally, some directions and insights for future work are detailed.