Abstract
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.
Similar content being viewed by others
Notes
Shapely: https://pypi.org/project/Shapely/
References
Centers for Disease Control and Prevention (2006) Ohio Emergency Management Agency: http://www.ema.ohio.gov/
Ocotlan DP, Ruiz-Vanoye JA, Loranca BB, Fuentes-Penna A, Barrera-Camara RA (2014) A survey of transportation problems
Kalcsics J (2015) Districting Problems, vol 1. Springer International Publishing, Cham, pp 595–622. ISBN 978-3-319-13111-5. https://doi.org/10.1007/978-3-319-13111-5_23
Gwalani H, Mikler AR, Ramisetty-Mikler S, O’Neill M (2017) Collection and integration of multi-spatial and multi-type data for vulnerability analysis in emergency response plans. In: Wohlgemuth V, Fuchs-Kittowski F, Wittmann J (eds) Advances and new trends in environmental informatics. ISBN 978-3-319-44711-7. Springer International Publishing, Cham, pp 89–101
Rahman M, Chen N, Islam MM, Dewan A, Pourghasemi HR, Washakh RMA, Nepal N, Tian S, Faiz H, Alam M, Ahmed N (2021) Location-allocation modeling for emergency evacuation planning with GIS and remote sensing: A case study of Northeast Bangladesh . Geosci Front 12:101095. ISSN 1674-9871. https://www.sciencedirect.com/science/article/pii/S1674987120302279
Akiko M, Dewan A, Corner RJ (2015) Population evacuation: evaluating spatial distribution of flood shelters and vulnerable residential units in Dhaka with geographic information systems. Nat Hazards 78:1859–1862
Ohi SJ, Kim AM (2021) Identifying critical corridors during an area-wide disruption by evaluating network bottleneck capacity. Int J Disaster Risk Reduction 64:102487. ISSN 2212-4209. https://www.sciencedirect.com/science/article/pii/S2212420921004489
Karaye IM, Thompson C, Horney JA (2019) Evacuation shelter deficits for socially vulnerable texas residents during hurricane harvey. Health Serv Res Manag Epidemiology 6:2333392819848885. pMID: 31205979 (Preprint https://doi.org/10.1177/2333392819848885)
Diaz JA , Fernandez E (2006) ISSN 0377-2217 feature Cluster on Scatter Search Methods for Optimization. http://www.sciencedirect.com/science/article/pii/S0377221704005557. Eur J Oper Res 169:570–585
Lorena LA, Senne EL (2004) ISSN 0305-0548. http://www.sciencedirect.com/science/article/pii/S030505480300039X. Comput Oper Res 31:863–876
Mulvey JM, Beck MP (1984) Solving capacitated clustering problems. European J Oper Res 18:339–348. ISSN 0377-2217. http://www.sciencedirect.com/science/article/pii/0377221784901553
Osman IH, Christofides N (1994) Capacitated clustering problems by hybrid simulated annealing and tabu search. Int Trans Oper Res 1:317–336. ISSN 0969-6016. http://www.sciencedirect.com/science/article/pii/0969601694900329
Franca PM, Sosa NM, Pureza V (1999) An adaptive tabu search algorithm for the capacitated clustering problem. Int Trans Oper Res 6:665–678. ISSN 0969-6016. http://www.sciencedirect.com/science/article/pii/S0969601699000179
Bansal N, Krishnaswamy R, Saha B (2011) On capacitated set cover problems. In: Goldberg LA, Jansen K, Ravi R, Rolim JDP (eds) Approximation, randomization, and combinatorial optimization. algorithms and techniques. ISBN 978-3-642-22935-0. Springer, Berlin, pp 38–49
Verter V (2011) Uncapacitated and Capacitated Facility Location Problems. Springer, Boston, pp 25–37. ISBN 978-1-4419-7572-0
Moreno L, de Aragão MP, Uchoa E (2006) Column generation based heuristic for a helicopter routing problem. In: Àlvarez C, Serna M (eds) Experimental Algorithms. ISBN 978-3-540-34598-5. Springer, Berlin, pp 219–230
El-Sherbeny NA (2010) Vehicle routing with time windows: An overview of exact, heuristic and metaheuristic methods. J King Saud Univ - Sci 22:123–131. ISSN 1018-3647. http://www.sciencedirect.com/science/article/pii/S1018364710000297
Rostamian Delavar M, Hajiaghaei-Keshteli M, Molla-Alizadeh-Zavardehi S (2010) Genetic algorithms for coordinated scheduling of production and air transportation. Expert Syst Appl 37:8255–8266. ISSN 0957-4174. https://doi.org/10.1016/j.eswa.2010.05.060
Ghoseiri K, Ghannadpour SF (2010) A hybrid genetic algorithm for multi-depot homogenous locomotive assignment with time windows. Appl Soft Comput 10:53–65. ISSN 1568-4946. http://www.sciencedirect.com/science/article/pii/S1568494609000726
Cordeau JF, Laporte G, Mercier A (2001) A unified tabu search heuristic for vehicle routing problems with time windows. J Oper Res Soc 52:928–936. ISSN 1476-9360. https://doi.org/10.1057/palgrave.jors.2601163
Taillard E, Badeau P, Gendreau M, Guertin F, Potvin J Y (1997) A tabu search heuristic for the vehicle routing problem with soft time windows. Trans Sci 31:170–186. (Preprint https://doi.org/10.1287/trsc.31.2.170)
Chao IM (2002) A tabu search method for the truck and trailer routing problem. Comput Oper Res 29:33–51. ISSN 0305-0548. http://www.sciencedirect.com/science/article/pii/S0305054800000563
Lin S W, Yu V F, Chou S Y (2009) Solving the truck and trailer routing problem based on a simulated annealing heuristic. Comput Oper Res 36:1683–1692. ISSN 0305-0548 selected papers presented at the Tenth International Symposium on Locational Decisions (ISOLDE X). http://www.sciencedirect.com/science/article/pii/S0305054808000762
Arbelaitz O, Rodriguez C, Zamakola I (2001) Low cost parallel solutions for the vrptw optimization problem
Bozkaya B, Erkut E, Laporte G (2003) ISSN 0377-2217. http://www.sciencedirect.com/science/article/pii/S0377221701003800. European J Oper Res 144:12–26
George JA, Lamar BW, Wallace CA (1997) Political district determination using large-scale network optimization
Jimenez T, Mikler A R, Tiwari C (2012) A novel space partitioning algorithm to improve current practices in facility placement. IEEE Trans Syst Man Cybern Syst Hum 42:1194–1205
Duque JC, Ramos R, Surinach J (2007) Supervised regionalization methods: a survey. Int Reg Sci Rev 30:195–220. (Preprint https://doi.org/10.1177/0160017607301605)
Shirabe T (2009) Districting modeling with exact contiguity constraints. Environ Plann B: Plann Des 36:1053–1066. (Preprint https://doi.org/10.1068/b34104)
Butsch A, Kalcsics J, Laporte G (2014) Districting for arc routing. INFORMS J Comput 26:809–824. (Preprint https://doi.org/10.1287/ijoc.2014.0600)
Benzarti E, Sahin E, Dallery Y (2013) Operations management applied to home care services: Analysis of the districting problem, vol 55. ISSN 0167-9236 1. Analytics and Modeling for Better HealthCare 2. Decision Making in Healthcare. http://www.sciencedirect.com/science/article/pii/S0167923612002709
D’Amico SJ, Wang SJ, Batta R, Rump CM (2002) A simulated annealing approach to police district design. Comput Oper Res 29:667–684. ISSN 0305-0548 location Analysis. http://www.sciencedirect.com/science/article/pii/S0305054801000569
Blais M, Lapierre SD, Laporte G (2003) Solving a home-care districting problem in an urban setting. J Oper Res Soc 54:1141–1147. ISSN 1476-9360 https://doi.org/10.1057/palgrave.jors.2601625
Baker JR, Clayton ER, Moore LJ (1989) Redesign of primary response areas for county ambulance services. European J Oper Res 41:23–32. ISSN 0377-2217 http://www.sciencedirect.com/science/article/pii/0377221789900350
Ferland JA , Guenette G (1990) Decision support system for the school districting problem. Oper Res 38:15–21. (Preprint https://doi.org/10.1287/opre.38.1.15)
Schoepfle O, Church RL (1991) A new network representation of a “classic” school districting problem, vol 25. ISSN 0038-0121. http://www.sciencedirect.com/science/article/pii/003801219190017L
Ricca F, Simeone B (2008) Local search algorithms for political districting. European J Oper Res 189:1409–1426. ISSN 0377-2217 http://www.sciencedirect.com/science/article/pii/S0377221707006273
Schwartzberg JE (1966) Reapportionment, Gerrymanders, and the Notion of Compactness. Minnesota Law Review, 1701. https://scholarship.law.umn.edu/mlr/1701/?utm_source=scholarship.law.umn.edu
Reock EC (1961) A note: measuring compactness as a requirement of legislative apportionment. JSTOR 5:70–74. https://www.jstor.org/stable/2109043
Adhia A, Grozdanov D, Ramezanzadeh M, Fisher Z (2019) Measuring compactness. https://fisherzachary.github.io/public/r-output.html
Gwalani H (2019) Source code: Fixed faxility contiguity constrained optimization. https://github.com/harsha2412/FixedFacility
United States Census Bureau (2019) Tiger/line shapefiles. www.census.gov/geographies/mapping-files/time-series/geo/tiger-line-file.html
Social Explorer (2019) American Community Survey- 5 Year Estimates. https://www.socialexplorer.com/blog/category/data-releases
Ramsey P, Leslie M (2012) Introduction to postgis: Chapter 9 geometries. http://postgis.net/workshops/postgis-intro/geometries.html
Funding
This work was supported by grants from 1. National Institutes of Health: Grant Number: 2 R56 LM011647-03, Title: Minimizing Access Disparities in Bio- Emergency Response Planning and 2. Texas Department of State Health Services: Project Title: Development and deployment of computational methods to facilitate response planning for POD placement and distribution of Medical Counter Measures from Regional RSS sites to PODs in Texas DSHS Region 6/5S.
Author information
Authors and Affiliations
Contributions
1. Harsha Gwalani: Conceptualization of this study, Data curation, Methodology, Software, Writing – original draft, and Formal Analysis
2. Chetan Tiwari: Writing - Original draft & preparation, Supervision, and Writing – review & editing
3. Marty O’Neill II: Data curation, and Writing – review & editing
4. Armin R Mikler: Funding acquisition, Supervision, Project administration, and Writing - Original draft preparation
Corresponding author
Ethics declarations
Competing interests
The authors have no competing interests to declare that are relevant to the content of this article.
Additional information
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Appendices
Appendix : A
A.1 Forcing contiguity
A demand point is transferred from facility fj to facility fk only if the service area corresponding to fj is adjacent to the service area for fk. While this condition is a necessary condition for contiguity of service areas, it does not guarantee contiguity of the resulting service areas for both fj and fk. As can be seen in Fig. 12 the donating facility service area will lose contiguity if the shaded demand point is selected for transfer to another facility. Figure 13 shows an example where a facility service area (f2) becomes disconnected after a demand point is transferred to it (from f1). Hence, if the contiguity of service areas is a hard constraint, then while selecting a demand point for transfer this condition needs to be taken into account. If the selection of the demand point from fj that minimizes the cost of transfer results in non-contiguous service areas then the next closest demand point is selected until this condition is met. If no such demand point can be found, the path is discarded.
Each service area can either be of polygon type or multi-polygon type [44]. If the service area is a polygon, it signifies that it is a contiguous region, however, if it is a multi-polygon it may be a collection of connected or disconnected polygons. Figure 14 shows examples of both a multi-polygon with connected polygons (I) and a multi-polygon with disconnected polygons(II). The contiguity condition is checked by traversing the graph formed by the polygons in a multi-polygon service area if any polygon is unreachable, it shows that the service area is non-contiguous.
A.2 Computational complexity
Theorem A.1
The number of iterations required to balance the demand at facilities is less than the number of demand points.
Proof
To simplify the proof we assume that each demand point has one unit of demand, hence the total demand is n and the average demand is n/p. Only one unit of demand is transferred between any two facilities in each iteration, hence the source can only lose a unit and a sink can gain only one unit of demand in each iteration. To maximize the number of iterations, we examine the scenario when the demand is most unbalanced. Each facility will have to have at least one demand point assigned to it, as the demand point where the facility is located at, is assigned to the facility. To create the most unbalanced scenario, we assume that one of p facilities, \(f^{\prime }\) has been assigned the remaining demand points. Hence \(demand(f^{\prime }) = (n-p+1)\) while, the demand at all remaining facilities is equal to demand(fi) = 1, for all fi ∈ F, \(f_i \neq f^{\prime }\). Since only one demand point is transferred from a source (\(f^{\prime }\)) to a sink fj in each iteration, to achieve the balance at \(f^{\prime }\), number of required iterations is equal to (n − p + 1 − n/p) = n + 1 − (p + n/p). The transfer may be indirect. b = (n + 1 − (p + n/p)). The required number of iterations is zero if the demands are already balanced. Therefore, 0 ≤ b < n. The condition of achieving balance in the more realistic scenario where demands are not uniform across all demand points is less restrictive because the algorithm terminates with a non-zero balance as discussed in the previous section, hence requires even less number of iterations. Hence the bound holds for the general case.
Rights and permissions
About this article
Cite this article
Gwalani, H., Tiwari, C., O’Neill II, M. et al. Creating contiguous service areas around points of dispensing for resource distribution during bio-emergencies. Geoinformatica 27, 461–490 (2023). https://doi.org/10.1007/s10707-022-00462-5
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10707-022-00462-5