This thesis, developed in collaboration with Explora s.n.c.—an Italian research and statistical consulting firm specializing in healthcare data—proposes a strategic redesign of the Emergency Medical Services (EMS) system in the Autonomous Province of Bolzano. The research pro- motes a radical clinical and logistical paradigm shift. In the past, less than 25% of critical inter- ventions received advanced care, forcing basic ambulances to cover severe emergencies due to structural limits. The proposed model imposes a strict mandate: 100% of High-Acuity codes must receive ALS units (Advanced Life Support, staffed by doctors and nurses). To achieve this, the thesis structures three core methodological contributions. The first contribution is a data-driven modeling upgrade. Past company models depended upon Maximal Covering Location Problems (MCLP) using theoretical Haversine aerial dis- tances. The proposed Optimization-Simulation framework abandons this geometric abstrac- tion. It processes over 322,000 telemetric records to engineer true road response time matrices. These mathematical structures integrate ISTAT travel times and specific correction indices, re- flecting true road friction based on clinical acuity and daily time bands (morning, afternoon, night). The second contribution focuses on spatial optimization, leveraging data from 2022 to 2025 to select active departure stations. To solve this problem, we consider a heuristic approach based on a Genetic Algorithm. Although this meta-heuristic provides no mathematical guarantee of finding the global optimum, it converges to a reliable geographic layout within reasonable computational times. The mathematical objective avoids the absolute minimization of aver- age response times. The Fitness Function guides the optimization to maximize the percentage of global in-target interventions while minimizing the number of open stations. A specific penalty constraint discards redundant infrastructures to save operational costs. The algorithm evaluates scenarios for High-Acuity (e.g., 7 hospital hubs plus Community Healthcare Cen- ters or historical bases) and Low-Acuity events. For High-Acuity, the model enforces the Es- sential Levels of Assistance (LEA) targets of 8 minutes (urban) and 20 minutes (rural). For Low-Acuity, it imposes strict control thresholds of 30 minutes (Green codes) and 60 minutes (White codes). These parameters are more restrictive than the British NHS benchmarks (120 and 180 minutes), evaluating whether compliance exceeds 90%. The third contribution provides structural validation. The optimized layouts undergo a dynamic stress test through a Digital Twin built in SimPy, isolating data from the year 2025. The simulator calculates the maximum peak of concurrent interventions (the exact number of ambulances operating at the same time). This dynamic metric dictates the precise vehicle and personnel allocation required for each time band and season. In conclusion, the framework validates the proposed redesign through a complete numeri- cal comparison against the 2025 real historical baseline. The results demonstrate that the engi- neered scenarios bridge urban coverage deficits and achieve superior clinical compliance using a rationalized structural fleet, proving the resilience and efficiency of the new logistical model.
This thesis, developed in collaboration with Explora s.n.c.—an Italian research and statistical consulting firm specializing in healthcare data—proposes a strategic redesign of the Emergency Medical Services (EMS) system in the Autonomous Province of Bolzano. The research pro- motes a radical clinical and logistical paradigm shift. In the past, less than 25% of critical inter- ventions received advanced care, forcing basic ambulances to cover severe emergencies due to structural limits. The proposed model imposes a strict mandate: 100% of High-Acuity codes must receive ALS units (Advanced Life Support, staffed by doctors and nurses). To achieve this, the thesis structures three core methodological contributions. The first contribution is a data-driven modeling upgrade. Past company models depended upon Maximal Covering Location Problems (MCLP) using theoretical Haversine aerial dis- tances. The proposed Optimization-Simulation framework abandons this geometric abstrac- tion. It processes over 322,000 telemetric records to engineer true road response time matrices. These mathematical structures integrate ISTAT travel times and specific correction indices, re- flecting true road friction based on clinical acuity and daily time bands (morning, afternoon, night). The second contribution focuses on spatial optimization, leveraging data from 2022 to 2025 to select active departure stations. To solve this problem, we consider a heuristic approach based on a Genetic Algorithm. Although this meta-heuristic provides no mathematical guarantee of finding the global optimum, it converges to a reliable geographic layout within reasonable computational times. The mathematical objective avoids the absolute minimization of aver- age response times. The Fitness Function guides the optimization to maximize the percentage of global in-target interventions while minimizing the number of open stations. A specific penalty constraint discards redundant infrastructures to save operational costs. The algorithm evaluates scenarios for High-Acuity (e.g., 7 hospital hubs plus Community Healthcare Cen- ters or historical bases) and Low-Acuity events. For High-Acuity, the model enforces the Es- sential Levels of Assistance (LEA) targets of 8 minutes (urban) and 20 minutes (rural). For Low-Acuity, it imposes strict control thresholds of 30 minutes (Green codes) and 60 minutes (White codes). These parameters are more restrictive than the British NHS benchmarks (120 and 180 minutes), evaluating whether compliance exceeds 90%. The third contribution provides structural validation. The optimized layouts undergo a dynamic stress test through a Digital Twin built in SimPy, isolating data from the year 2025. The simulator calculates the maximum peak of concurrent interventions (the exact number of ambulances operating at the same time). This dynamic metric dictates the precise vehicle and personnel allocation required for each time band and season. In conclusion, the framework validates the proposed redesign through a complete numeri- cal comparison against the 2025 real historical baseline. The results demonstrate that the engi- neered scenarios bridge urban coverage deficits and achieve superior clinical compliance using a rationalized structural fleet, proving the resilience and efficiency of the new logistical model.
Data-Driven Simulation-Optimization for Alpine EMS Logistics: The Case of Bolzano–South Tyrol
LOVATO, EDOARDO
2025/2026
Abstract
This thesis, developed in collaboration with Explora s.n.c.—an Italian research and statistical consulting firm specializing in healthcare data—proposes a strategic redesign of the Emergency Medical Services (EMS) system in the Autonomous Province of Bolzano. The research pro- motes a radical clinical and logistical paradigm shift. In the past, less than 25% of critical inter- ventions received advanced care, forcing basic ambulances to cover severe emergencies due to structural limits. The proposed model imposes a strict mandate: 100% of High-Acuity codes must receive ALS units (Advanced Life Support, staffed by doctors and nurses). To achieve this, the thesis structures three core methodological contributions. The first contribution is a data-driven modeling upgrade. Past company models depended upon Maximal Covering Location Problems (MCLP) using theoretical Haversine aerial dis- tances. The proposed Optimization-Simulation framework abandons this geometric abstrac- tion. It processes over 322,000 telemetric records to engineer true road response time matrices. These mathematical structures integrate ISTAT travel times and specific correction indices, re- flecting true road friction based on clinical acuity and daily time bands (morning, afternoon, night). The second contribution focuses on spatial optimization, leveraging data from 2022 to 2025 to select active departure stations. To solve this problem, we consider a heuristic approach based on a Genetic Algorithm. Although this meta-heuristic provides no mathematical guarantee of finding the global optimum, it converges to a reliable geographic layout within reasonable computational times. The mathematical objective avoids the absolute minimization of aver- age response times. The Fitness Function guides the optimization to maximize the percentage of global in-target interventions while minimizing the number of open stations. A specific penalty constraint discards redundant infrastructures to save operational costs. The algorithm evaluates scenarios for High-Acuity (e.g., 7 hospital hubs plus Community Healthcare Cen- ters or historical bases) and Low-Acuity events. For High-Acuity, the model enforces the Es- sential Levels of Assistance (LEA) targets of 8 minutes (urban) and 20 minutes (rural). For Low-Acuity, it imposes strict control thresholds of 30 minutes (Green codes) and 60 minutes (White codes). These parameters are more restrictive than the British NHS benchmarks (120 and 180 minutes), evaluating whether compliance exceeds 90%. The third contribution provides structural validation. The optimized layouts undergo a dynamic stress test through a Digital Twin built in SimPy, isolating data from the year 2025. The simulator calculates the maximum peak of concurrent interventions (the exact number of ambulances operating at the same time). This dynamic metric dictates the precise vehicle and personnel allocation required for each time band and season. In conclusion, the framework validates the proposed redesign through a complete numeri- cal comparison against the 2025 real historical baseline. The results demonstrate that the engi- neered scenarios bridge urban coverage deficits and achieve superior clinical compliance using a rationalized structural fleet, proving the resilience and efficiency of the new logistical model.| File | Dimensione | Formato | |
|---|---|---|---|
|
Lovato_Edoardo.pdf
Accesso riservato
Dimensione
6.93 MB
Formato
Adobe PDF
|
6.93 MB | Adobe PDF |
The text of this website © Università degli studi di Padova. Full Text are published under a non-exclusive license. Metadata are under a CC0 License
https://hdl.handle.net/20.500.12608/115843