A Realistic Discrete Event Simulation model for Ambulance Location and Deployment within a regional Emergency Medical Service

arXiv:2604.24394 · eess.SY, cs.SY · Submitted 2026-04-27 · Read on arXiv

Listen

Radio episode about this paper

Transcript

Introduction to the show: ident: Robotics Radio. Generated commentary on the latest robotics and control papers.

Rosa: We're looking at the paper: "A Realistic Discrete Event Simulation model for Ambulance Location and Deployment within a regional Emergency Medical Service".

Dev: Here's the gist of the paper. The paper presents a general Discrete Event Simulation (DES) model designed to capture the stochastic behavior and workflow of regional ambulance emergency systems, which is important because it allows managers to improve effectiveness and quickness of the entire regional EMS system.

Rosa: System Dynamics Modeling The proposed model aims to reproduce the operational dynamics of an ambulance emergency system by explicitly modeling the full operational path of an ambulance mission.

Dev: The model incorporates several features that enhance realism compared to existing literature, such as incorporating operational delays like telephone triage and ambulance preparation times using empirical distributions estimated from historical data.

Rosa: Furthermore, the model considers times due to ambulance offload delay and ambulance ramping, which are crucial for monitoring how long ambulances remain unavailable while waiting at the ED.

Title and authors: Rosa: So we're looking at a paper called "A Realistic Discrete Event Simulation model for Ambulance Location and Deployment within a regional Emergency Medical Service" by De Santisa and the team. It sounds like they built a simulation to really map out how ambulances are placed and sent out in an emergency medical service.

Dev: Yeah, the title makes it clear they're focusing on making it realistic, not just some abstract math exercise. They’re trying to capture the actual flow of an ambulance mission across a whole region, which is important because EMS effectiveness really depends on where those ambulances are stationed and how they get to the patient fast.

Taro: I wonder if a simulation can actually get that complex? Usually, these models simplify things too much, so I'm curious if this one manages to model the real mess of an emergency system.

Rosa: Exactly, because as you know with these models, if it doesn't capture the delays or the actual decision-making process well, you end up with a model that doesn't help managers make good choices about resource placement.

Dev: That’s the key difference here; they are explicitly modeling the full operational path of an ambulance mission, which is a big step because it means they aren't just looking at where the ambulance starts and ends, but everything in between.

The paper's summary: Rosa: The paper says their goal is to reproduce the stochastic behavior of these systems, meaning they’re trying to mimic how random events happen in an EMS operation so managers can improve how fast and effective the whole regional system is.

Dev: It does that by incorporating several features that make it much more realistic than what's out there now. They are adding things like operational delays, such as telephone triage and the time it takes to get an ambulance prepared for a call, which they model using distributions from historical data.

Taro: Wait, so they’re modeling those preparation times based on real historical stuff? That’s interesting because in the real world, that prep time is totally unpredictable and depends on what kind of call it is.

Rosa: Right, and they also look at delays related to when an ambulance has to wait after an offload at the emergency department, and they model ambulance ramping too—that’s how long ambulances sit idle waiting for more calls or patients.

Dev: Plus, they are using a more sophisticated way to decide which emergency department an ambulance should go to based on what the patient actually has, instead of just looking at how close it is geographically.

The paper's improvements: Rosa: They point out that existing models often rely on simplifying assumptions that limit their ability to represent real EMS operations, and this paper directly addresses those limitations by weakening some of those assumptions.

Dev: Specifically, they incorporate operational delays like telephone triage and ambulance preparation times using empirical distributions estimated from historical data, which is a big step because most literature neglects those kinds of bottlenecks.

Taro: And they also consider the time due to ambulance offload delay and ambulance ramping, which are crucial for monitoring how long ambulances stay unavailable while waiting at the emergency department, so it gives a better picture of system strain.

Rosa: On top of that, they use these more detailed criteria for choosing the emergency department based on the diagnosed pathology instead of just using simple proximity rules. That adds a layer of realism to where an ambulance actually goes after arrival.

Conclusion: Dev: So, to wrap up, this model gives managers a flexible and reusable tool for planning and resource allocation because it accounts for the demand concentration, temporal availability of resources, and how spatial positioning interacts with these random system dynamics.

Rosa: It confirms that when you make deployment decisions, you really need to think about where the demand is concentrated in space and time, not just where it is geographically close to a base.

Taro: For me, the most important finding is that structural capacity reinforcement at a central high-demand location actually produces robust improvements across all response time thresholds; specifically, adding one H24 ambulance at the Rieti fire station showed statistical dominance in their scenario analysis.

Dev: That’s significant evidence showing that targeted deployment decisions based on demand concentration and temporal availability yield the most substantial gains in system performance, which is what this paper confirms about the realistic discrete event simulation model for Ambulance Location and Deployment within a regional Emergency Medical Service.

Dipartimento di Ingegneria Informatica, Automatica e Gestionale “A. Ruberti SAPIENZA - Universitaè di Roma via Ariosto 25 Roma 00185 Italy · Dipartimento di Biomedicina e Prevenzione Universitaè di Roma Tor Vergata Viale Montpellier 1 Roma 00133 Italy · Azienda Regionale Emergenza Sanitaria ARES 118 Via Portuense 240 Roma 00149 Italy

eess.SY, cs.SY

Submitted: 2026-04-27

Updated: 2026-11-06

Comments: 41 pages, 12 figures

License: http://creativecommons.org/licenses/by/4.0/

Importance score: 88/100

The gist: The paper presents a general Discrete Event Simulation (DES) model designed to capture the stochastic behavior and workflow of regional ambulance emergency systems…

Key concepts

Discrete Event Simulation (DES)
A modeling technique used to simulate systems where events happen at specific points in time. In this context, it tracks the journey of emergency calls and ambulances through various stages, allowing managers to observe how the entire regional EMS workflow behaves realistically under different conditions.
Operational Delays (e.g., Triage/Preparation)
These are realistic time buffers added to the model for actions that occur before an ambulance can respond. The model uses historical data distributions to estimate these delays, ensuring that the simulation accurately reflects the time ambulances spend on non-emergency tasks like telephone triage and preparing for a call.
Scenario Analysis
This is a testing method where managers change specific system parameters—like adding an ambulance or changing deployment locations—to see the resulting impact on key performance indicators, such as response time coverage. It helps identify the most effective strategic changes for resource allocation.

Terminology

Summary

The paper presents a general Discrete Event Simulation (DES) model designed to capture the stochastic behavior and workflow of regional ambulance emergency systems, which is important because it allows managers to improve effectiveness and quickness of the entire regional EMS system. p. 0

Where the paper says it

This paper presents a general Discrete Event Simulation (DES) model designed to capture the stochastic behaviour and workflow of regional ambulance emergency systems. The proposed model incorporates and integrates information collected from different sources, reproducing very accurately the operation of the ambulance system, thus allowing a more comprehensive and realistic analysis.

System Dynamics Modeling

The proposed model aims to reproduce the operational dynamics of an ambulance emergency system by explicitly modeling the full operational path of an ambulance mission. The model incorporates several features that enhance realism compared to existing literature, such as incorporating operational delays like telephone triage and ambulance preparation times using empirical distributions estimated from historical data. p. 2

Where the paper says it

In particular, this is achieved by weakening some simplifying assumptions commonly adopted in the literature. For instance, we incorporate operational delays that are frequently neglected in the literature [26, 68], such as telephone triage and ambulance preparation times, which are modeled using empirical distributions estimated from historical data.

Furthermore, the model considers times due to ambulance offload delay and ambulance ramping, which are crucial for monitoring how long ambulances remain unavailable while waiting at the ED. p. 2
Where the paper says it

Hence, healthcare professionals must stay with the patient (sometimes for hours), keeping the ambulance unavailable for new emergencies. Since this phenomenon occurs very frequently, it is important to consider this time separately, in order to monitor how long ambulances remain unavailable while waiting at the ED.

The model also uses more sophisticated criteria for choosing the ED based on diagnosed pathology rather than just proximity. p. 2
Where the paper says it

Another model enhancement concerns the use of more sophisticated criteria for choosing the ED where the patient is delivered. In fact, in our model, the choice is not only simply based on proximity criteria (typically the nearest ED) [39], but on the diagnosed pathology, as actually performed in practice.

Model Structure and Assumptions

The general purpose and reusable DES model is structured to reflect the flowchart of the ambulance emergency system processes reported in Figure 1. The model entities are represented by emergency calls (temporary entities) and ambulances (permanent entities). p. 0

Where the paper says it

, locations where ambulances and their crews are strategically positioned, ready to respond to emergency calls. This paper presents a general Discrete Event Simulation (DES) model designed to capture the stochastic behaviour and workflow of regional ambulance emergency systems.

Key assumptions adopted in the model include that restocking, crew shifts and maintenance are not considered. p. 8
Where the paper says it

Assumptions We now itemize the assumptions adopted in our model. Most of them are common to the models already proposed in literature: • restocking, crew shifts and maintenance are not considered; • travel times are assumed to be deterministic, but a detailed time-varying travel is used, depending upon time slot of the day and urgency level; • the ambulance diversion is not allowed, i.

Travel times are assumed to be deterministic, but a detailed time-varying travel is used depending on the time slot of the day and urgency level. p. 8
Where the paper says it

Assumptions We now itemize the assumptions adopted in our model. Most of them are common to the models already proposed in literature: • restocking, crew shifts and maintenance are not considered; • travel times are assumed to be deterministic, but a detailed time-varying travel is used, depending upon time slot of the day and urgency level; • the ambulance diversion is not allowed, i.

The model also assumes that on-idle redispatch is not allowed, meaning an ambulance assigned to an emergency call can not be changed since another ambulance closer to the location of the call has become available. p. 4
Where the paper says it

Unfortunately, these models do not take into account possible unavailability of a vehicle of responding to a call since already busy in a previous task. Therefore subsequent extensions introduced stochastic models considering the probability that an ambulance is available on emergency call arrival.

Data Inputs and Calibration

Accurate data collection is fundamental for building a valid model, requiring inputs such as area of interest and characteristics, geographical locations of bases and EDs, inter-zone travel times, call arrivals with spatial and temporal indications, dispatching rules, and times to perform tasks. Travel times are computed using routing services like OpenRouteService API to account for the actual road network configuration rather than simple Euclidean distances. p. 2

Where the paper says it

Furthermore, we adopt a flexible demand generation mechanism with spatial hierarchy, capturing both temporal variability and fine-grained spatial heterogeneity of emergency calls. Finally, travel times are computed on the actual road network and calibrated using historical observations, thus avoiding the common simplifying assumptions of Euclidean distances or constant speeds [49, 68].

A multiplicative calibration factor is introduced to adjust routing service travel times based on time slot and urgency class, resulting in a calibrated travel time used by the simulation model. p. 8
Where the paper says it

Assumptions We now itemize the assumptions adopted in our model. Most of them are common to the models already proposed in literature: • restocking, crew shifts and maintenance are not considered; • travel times are assumed to be deterministic, but a detailed time-varying travel is used, depending upon time slot of the day and urgency level; • the ambulance diversion is not allowed, i.

Service times like telephone triage and ambulance preparation are modeled using probability distributions estimated from historical data, distinguishing between emergency cases and non-emergency cases. p. 2
Where the paper says it

In particular, this is achieved by weakening some simplifying assumptions commonly adopted in the literature. For instance, we incorporate operational delays that are frequently neglected in the literature [26, 68], such as telephone triage and ambulance preparation times, which are modeled using empirical distributions estimated from historical data.

Validation and Scenario Analysis

The model was validated using a real-world case study provided by ARES 118 concerning the Lazio region of Italy. Validation results showed a close agreement between simulated and historical data in terms of annual call volumes, response-time coverage across multiple thresholds, and ambulance utilization at the base level. p. 3

Where the paper says it

Validation results showed a close agreement between simulated and real system, so that the model can be successfully adopted as an accurate representation of the EMS under study. The scenario analysis performed on the real case showed that decisions concerning ambulance deployment on a territory should explicitly account for demand concentration, temporal availability of resources, and the interaction between spatial positioning and stochastic system dynamics.

Scenario analysis explores how changes in resource availability and spatial deployment affect system performance, focusing on the coverage KPI. p. 3
Where the paper says it

Validation results showed a close agreement between simulated and real system, so that the model can be successfully adopted as an accurate representation of the EMS under study. The scenario analysis performed on the real case showed that decisions concerning ambulance deployment on a territory should explicitly account for demand concentration, temporal availability of resources, and the interaction between spatial positioning and stochastic system dynamics.

The results indicate that structural capacity reinforcement at a central high-demand location produces robust, system-wide improvements across all RT thresholds. Specifically, Scenario 5, corresponding to the addition of one H24 ambulance at the Rieti fire station, exhibits statistical dominance. p. 8
Where the paper says it

Furthermore, ambulances may operate under different shift schedules, typically H12, corresponding to daytime coverage (12 hours), and H24, corresponding to continuous dayand-night coverage (24 hours). Consequently, the effective availability of ambulances in the EMS system depends not only on the number of vehicles but also on the shift schedule assigned to each ambulance.

Conclusion and Future Directions

The proposed DES framework provides a flexible and reusable tool for improving EMS planning and resource allocation. The study confirms that deployment decisions should explicitly account for demand concentration, temporal availability of resources, and the interaction between spatial positioning and stochastic system dynamics. p. 3

Where the paper says it

Validation results showed a close agreement between simulated and real system, so that the model can be successfully adopted as an accurate representation of the EMS under study. The scenario analysis performed on the real case showed that decisions concerning ambulance deployment on a territory should explicitly account for demand concentration, temporal availability of resources, and the interaction between spatial positioning and stochastic system dynamics.

Future research directions include the integration of demand forecasting models to enable proactive and adaptive deployment policies. p. 6
Where the paper says it

Finally, we mention how the rapid development of Artificial Intelligence has recently opened new perspectives for EMS decision support systems. Artificial Intelligence techniques can enhance both situational awareness and proactive decision making by enabling real-time monitoring, anomaly detection, and demand forecasting.

The framework's key strength is its full scalability and adaptability, enabling it to effectively accommodate varying system sizes, demand levels, and operational complexities.

Pruning redundant reasoning tokens

The paper presents a general Discrete Event Simulation (DES) model designed to capture the stochastic behavior and workflow of regional ambulance emergency systems, which is important because it allows managers to improve effectiveness and quickness of the entire regional EMS system.

How the training signal is built

The call generation process follows a two-stage mechanism where a call arrival time is drawn from the interarrival distribution of a certain call generation zone, and then within that zone, a specific call square is randomly selected. The temporal component of call generation is modeled through an analysis of historical demand patterns, retaining only those temporal variables that show a statistically significant influence on call frequency. p. 2

Where the paper says it

Furthermore, we adopt a flexible demand generation mechanism with spatial hierarchy, capturing both temporal variability and fine-grained spatial heterogeneity of emergency calls. Finally, travel times are computed on the actual road network and calibrated using historical observations, thus avoiding the common simplifying assumptions of Euclidean distances or constant speeds [49, 68].

The spatial component is modelled using a hierarchical structure subdividing the area under study into generation zones and smaller call squares. p. 2
Where the paper says it

Furthermore, we adopt a flexible demand generation mechanism with spatial hierarchy, capturing both temporal variability and fine-grained spatial heterogeneity of emergency calls. Finally, travel times are computed on the actual road network and calibrated using historical observations, thus avoiding the common simplifying assumptions of Euclidean distances or constant speeds [49, 68].

Where the gains come from

The scenario analysis demonstrates that increasing ambulance availability in high-demand and strategically central locations yields the most significant and robust improvements in system performance. The addition of a full-time (H24) ambulance at the Rieti fire station produces the largest gains in coverage for both urgent and non-urgent calls, reflecting the strong concentration of population and call volume in that area. p. 3

Where the paper says it

To show the reliability of the proposed general model, we cosidered a real case provided by the Azienda Regionale Emergenza Sanitaria - ARES 118 concerning a territory within the Lazio region of Italy, including the medium-size city of Rieti along with its province which also consists of sparsely populated areas. The rationale behind this choice is to consider a wide area with different population density so that the emergency calls are expected to be variously distributed.

This confirms that deployment decisions should explicitly account for demand concentration, temporal availability of resources, and the interaction between spatial positioning and stochastic system dynamics. p. 3
Where the paper says it

Validation results showed a close agreement between simulated and real system, so that the model can be successfully adopted as an accurate representation of the EMS under study. The scenario analysis performed on the real case showed that decisions concerning ambulance deployment on a territory should explicitly account for demand concentration, temporal availability of resources, and the interaction between spatial positioning and stochastic system dynamics.

Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.

Improvements for AI systems

  1. Improved demand forecasting enables proactive deployment by using predictive analytics thus facilitates the dynamic pre-deployment of ambulances, reducing response times and improving patient outcomes [31]. This allows the AI system to anticipate future demand patterns based on historical data and adjust ambulance placement before emergency calls arrive.

  2. Enhanced real-time monitoring and anomaly detection allow for proactive decision making by exploiting vast amounts of historical and real-time data (e.g., travel times, hospital occupancy, etc.) that can be exploited to predict future demand [19, 33, 38, 42, 47]. The AI can detect unexpected surges or bottlenecks in the system instantly.

  3. Integration within a simulation-optimization framework enables the system to systematically search for optimal fleet compositions and base configurations under multiple and possibly conflicting objectives [74]. This allows the AI to move beyond static models by iteratively testing how different ambulance deployments affect various KPIs like response time coverage across different demand scenarios.

  4. The model can accurately capture operational complexities such as Ambulance Offload Delay (AOD) and ambulance ramping, ensuring that the simulation reflects the reality where healthcare professionals must stay with the patient (sometimes for hours), keeping the ambulance unavailable for new emergencies [41]. This leads to more realistic predictions of system strain during high-demand periods.

  5. The system can use validated scenario analysis to test strategic changes, such as addition of one H24 ambulance at the Rieti fire station, and quantify their impact on coverage, confirming that structural capacity reinforcement at a central high-demand location produces robust, system-wide improvements [33]. This provides managers with evidence-based recommendations for resource allocation.

Abstract

The objective of Emergency Medical Services (EMSs) is to promptly respond to calls from citizens for first aid, providing pre-hospital care and, if necessary, to transfer patients to an appropriate Emergency Department (ED) by ambulance. The efficiency of such a system strongly depends on the deployment of ambulance home bases, i.e., locations where ambulances and their crews are strategically positioned, ready to respond to emergency calls. This paper presents a general Discrete Event Simulation (DES) model designed to capture the stochastic behaviour and workflow of regional ambulance emergency systems. The proposed model incorporates and integrates information collected from different sources, reproducing very accurately the operation of the ambulance system, thus allowing a more comprehensive and realistic analysis. To show the applicability and reliability of the proposed general model, a case study provided by the Azienda Regionale Emergenza Sanitaria - ARES 118 (an Italian Regional Emergency Medical Services Authority - ARES 118) is presented. It concerns a territory within the Lazio region of Italy, including a medium-size city along with sparsely populated areas. The reported results about scenario analyses highlight how the model we propose can be fruitfully used by the managers to improve effectiveness and quickness of the entire regional EMS system.

Related papers