Dynamic, robust models to quantify the impact of decentralization in post-disaster health care facility location decisions

We consider post-disaster health care facility location problems in which the need for facilities is not known until post-disaster, the need changes spatially and temporally as the response unfolds, and multiple organizations wish to open and close facilities over short time horizons based on dynamic needs and resource levels. In this context, accessibility, or the opportunity for individuals to receive services provided by facilities, and equity, or fairness, among individuals are both important. Accessibility and equity can be negatively impacted when facility location decisions are not coordinated among responding organizations, but coordination is difficult in large-scale humanitarian response efforts. In this paper, we introduce a methodology to quantify the impact on accessibility and equity resulting from lack of coordination in locating post-disaster health care service facilities. Our method includes a dynamic, scenario-based, robust optimization model to choose facility locations that optimize accessibility while ensuring equity. This hypothetical centralized solution may then be compared with actual uncoordinated facility location decisions to identify opportunities for improvement and build support for enhanced coordination. We demonstrate our approach using data from the international response to the 2010 Haiti cholera outbreak.