LOCATION OPTIMIZATION AND SIMULATION FOR THE ANALYSIS OF EMERGENCY MEDICAL SERVICE SYSTEMS (FACILITY, BALTIMORE, MARYLAND).

摘要

Provision of effective emergency medical service (EMS) is a task faced by virtually all municipalities. Since spatial and temporal demand changes must inevitably alter the quality of service in urban areas, there must be an on-going response to these changes. Typically the response is relocation of some ambulances or location of additional ambulances, depending upon budgetary constraints. These long-range location decisions are complicated by two primary EMS system features: (1) the conflicting multiple objectives inherent in the decision-making process and (2) the complex spatial, temporal and stochastic nature of demand and supply for ambulance service.; A methodological approach using an optimization model, the P-median Transportation problem (PMTP), and simulation was developed. The methodology was applied to the relocation of up to six ambulances (the maximum that can be relocated to advantage) in Baltimore, Maryland. The results show that the methodology improves our ability to capture the two primary, complicating features of EMS systems.; First, the PMTP is a multiobjective model. The primary objective of the PMTP is the same as the primary performance criterion of an EMS system--overall mean response time. However, there are other dimensions of the quality of service in EMS systems. A second objective of balancing ambulance utilization is incorporated into the PMTP in the form of maximum workload constraints. Workload is a valid and recently much publicized measure of EMS performance in so far as it translates into quality of service; overburdened ambulance personnel may not perform efficiently. Spatial equity concerns can also be reflected in the PMTP.; Second, the stochastic and dynamic nature of EMS systems are also dealt with using the PMTP. At present, there is no definitive, general optimization model that accurately depicts EMS systems behavior. Several models have been formulated to address the difficulties of stochastic demand, ambulance availability and nonclosest demand-facility assignment. Few have been tested, however. Solutions generated from the PMTP are tested with simulation and are compared with solutions from the P-median problem (PMP), a forerunner of the PMTP. The results indicate that the PMTP improves the representation of the dynamics and stochasticity of the EMS system.; The importance of using simulation both to test optimization models and to verify location solutions generated from these models is also explored. The results show that most mean response time reductions predicted by optimization are not statistically significant in the simulated system. Moreover, some solutions obtained from the PMP erroneously predict response time reductions; the converse occurs. Many of the results and observations may be generalizable to other EMS systems and other uses of optimization and simulation.