Institute of Management ifu, Ruhr University Bochum, Universitätsstraße 150, 44801, Bochum, Germany.
Chair of Operations Management, RWTH Aachen University, Kackertstr. 7, 52072, Aachen, Germany.
Health Care Manag Sci. 2022 Mar;25(1):42-62. doi: 10.1007/s10729-021-09562-4. Epub 2021 Jul 13.
In order to allocate limited resources in emergency medical services (EMS) networks, mathematical models are used to select sites and their capacities. Many existing standard models are based on simplifying assumptions, including site independency and a similar system-wide busyness of ambulances. In practice, when a site is busy, a call is forwarded to another site. Thus, the busyness of each site depends not only on the rate of calls in the surrounding area, but also on interactions with other facilities. If the demand varies across the urban area, assuming an average system-wide server busy fraction may lead to an overestimation of the actual coverage. We show that site interdependencies can be integrated into the well-known Maximum Expected Covering Location Problem (MEXCLP) by introducing an upper bound for the busyness of each site. We apply our new mathematical formulation to the case of a local EMS provider. To evaluate the solution quality, we use a discrete event simulation based on anonymized real-world call data. Results of our simulation-optimization approach indicate that the coverage can be improved in most cases by taking site interdependencies into account, leading to an improved ambulance allocation and a faster emergency care.
为了在紧急医疗服务（EMS）网络中分配有限的资源，数学模型被用于选择站点及其容量。许多现有的标准模型都是基于简化的假设，包括站点独立性和救护车在整个系统中的相似繁忙程度。实际上，当一个站点繁忙时，呼叫会被转发到另一个站点。因此，每个站点的繁忙程度不仅取决于其周边地区的呼叫率，还取决于与其他设施的交互。如果需求在城市区域内存在差异，假设整个系统范围内的服务器繁忙程度平均值可能会导致对实际覆盖范围的高估。我们通过为每个站点的繁忙程度引入一个上限，将站点之间的相互依赖性整合到著名的最大期望覆盖位置问题（MEXCLP）中。我们将新的数学公式应用于当地 EMS 提供商的案例。为了评估解决方案的质量，我们使用基于匿名真实世界呼叫数据的离散事件模拟。我们的仿真优化方法的结果表明，在大多数情况下，考虑站点之间的相互依赖性可以提高覆盖范围，从而改善救护车的分配和更快的紧急护理。
