Robust facility location of container clinics: A South African application

Abstract

There is a lack of dynamic facility location models for developing countries that consider the changes in the problem environment over time, such as patient population and population migration. Therefore, this paper focuses on using optimization and goal programming to locate health care facilities in an uncertain environment using multiple possible future urban development senarios. To achieve this, a robust multi-objective facility location model is developed and used to determine locations for container clinic deployment over multiple years in selected communities in South Africa. A synthetic population and urban growth simulation model are used to estimate population density and distribution from 2018 to 2030 for three development senarios. The results from the urban growth simulation model are then used as input into the facility location model to locate facilities whilst considering the three future development scenarios. Results of the model indicate that the robust model can be used to find locations that provide a relatively good solution to all considered development scenarios, providing key role players with quantitative decision support during network design under uncertainty. An accessibility analysis investigates the impact of the prescribed accessibility percentage on model results and a budget analysis evaluates the impact of a case that includes a budget constraint. From these two analyses it is illustrated that the model is sensitive to changes in parameters and that the model can be used by key stakeholders to combine network design and urban development planning for improved decision making.