Creating contiguous service areas around points of dispensing for resource distribution during bio-emergencies

• Published in: GeoInformatica 2023-07, Vol.27 (3), p.461-490
• Main Authors: Gwalani, Harsha, Tiwari, Chetan, O’Neill II, Marty, Mikler, Armin R
• Format: Article
• Language: English
• Subjects: Algorithms; Catchment area; Catchment areas; Combinatorial analysis; Computer Science; Cost function; Data Structures and Information Theory; Decision support systems; Demand; Dispensing; Distance; Distribution; Emergencies; Emergency plans; Emergency response; Geographical Information Systems/Cartography; Greedy algorithms; Heuristic methods; Industrial locations; Information Storage and Retrieval; Multimedia Information Systems; Optimization; Populations; Public health; Service areas
• ISSN: 1384-6175; 1573-7624
• DOI: 10.1007/s10707-022-00462-5

Response plans in preparation for public health emergencies often involve the setup of facilities like shelters, ad-hoc clinics, etc. to serve the affected population. While public health authorities frequently have prospective facility locations, balancing the demand or population at these facilities can be challenging. Assigning populations to their closest facilities may lead to uneven distribution of demand. This research proposes a novel greedy heuristic algorithm to create service areas around given facilities such that the population to be served by each facility is uniform or proportional to available resources. This algorithm has been implemented in the context of response plans for bio-emergencies in Denton County, Texas, USA. Given the location of Points of Dispensing (PODs), the objective is to create contiguous catchment areas, each served by one POD such that demand distribution constraints are satisfied. While the demand distribution constraints are hard constraints, it is also preferred that populations are mapped to PODs as close to them as possible. A response plan defines a mapping of populations to facilities and presents a combinatorial optimization problem in which the average distance between population locations and PODs is the cost function value, and demand equity and contiguity of catchment areas are hard constraints. We present a decision support system for planners to select solutions based on the compactness of catchment areas, the average distance between populations and PODs, and execution time, given that all solutions have contiguous catchment areas and balanced demand.