The talk introduces a new approach to solve multifacility location problems based on mixed integer programming and algorithms for minimizing differences of convex (DC) functions. This class of multifacility location problems is very difficult to solve because of its intrinsic discrete, nonconvex, and nondifferentiable nature. We first reformulate the problem under consideration as a continuous optimization problem then develop a new DC type algorithm involving Nesterov's smoothing. We also implement our method with MATLAB, numerical tests are done on both artificial and real data sets.
Applications of DCA in Solving Multifacility Location Problems Based on Mixed Integer Programming
Tuyen Tran, Loyola University ChicagoAuthors: Tuyen Tran, Anuj Bajaj, Mau Nam Nguyen, Boris Mordukhovich
2023 AWM Research Symposium
Women in Tensor Optimization [Organized by Longxiu Huang and Jing Qin]