A Lagrangian Heuristic for a Variant of Capacitated Facility Location with Single Source Constarints

Abstract

Facility location problems (FLP) are extensively studied in the literature in the context of supply chain management. Wide variety of real life situations are analyzed and modeled using techniques developed for FLP. In this thesis we take a comparably new model, Capacitated Facility Location with Single Source constraints (CFLPSS) from the literature and add an additional feature of Minimum Supply (MM) requirements (CFLPSSMM). Then we devise a Lagrangian Heuristic, which is highly efficient for CFLPSS models and for this new variant of CFLPSS. This heuristic, which is modified from the heuristics devised for CFLPSS, is then tested both on data from the literature and on new data set. Results indicate that it can be a resourceful alternative; especially the lower bounds provided by the heuristic are quite effective both for CFLPSS and CFLPSSMM.