Hybrid heuristic for dynamic location-allocation on micro-credit territory design

Abstract

This paper presents a two phase mixedinteger program for the commercial territory designproblem of a micro financing institution. After thelocations of the territory centers are determined, thecustomer allocation is done with respect to suchplanning criteria as total workload, amount of loans, andprofit allocation. In order to solve this model for largeinstances, we propose a hybrid heuristic that includesfixing variables, perturbation analysis, and dynamicrelocation of territory centers. We perform acomprehensive statistical analysis that provides novelinsights about the interplay of the heuristics in a largescale mixed integer program. The efficiency of the hybridheuristic is tested and its effectiveness to find nearoptimal solutions with a reasonably small computationaleffort is discussed.