Emergency Department Agent-Based Simulation

This project was developed as the third semester work for the Discrete Optimization course at the Faculty of Management Science and Informatics (FRI), University of Žilina.

The primary objective is an agent-based simulation study of a regional hospital's emergency department (ED). Facing limited personnel capacities and a growing number of patients, the simulation aims to determine the optimal structure and minimum number of medical staff (doctors and nurses) required to maintain efficient operations and minimize patient wait times.

• Multi-Agent Architecture: Built using the ABASim (Agent-Based Architecture Simulation) framework (v4.3+). The system utilizes a complex hierarchy of agents, including ModelAgent, UrgentAdmissionAgent, MovementAgent, and dedicated resource agents (NurseAgent, DoctorAgent, RoomAgent) to orchestrate the simulation.
• Data-Driven Distributions: Patient arrival rates were strictly modeled based on the analysis of provided historical data. Walk-in arrivals were fitted to an Exponential distribution, while ambulance arrivals were successfully fitted to a Beta distribution following Chi-square goodness-of-fit testing.
• Intricate Queuing Logic: The model distinguishes between walk-in patients and ambulance patients, automatically granting priority to ambulance arrivals.
• GUI & Animation: Includes a graphical user interface with live 2D animation tracking the movement of patients and medical staff through the hospital environment.

Simulation Resources

The simulation manages a continuous 24/7 operation over a 28-day replication period.

• Resources: The ED contains 12 examination rooms (5 of Type A, 7 of Type B), alongside a variable number of doctors and nurses.
• Stage 1 - Triage: Requires the simultaneous availability of a nurse and a Type B room. Patients are then assigned a medical priority from 1 (highest) to 5 (lowest).
• Stage 2 - Medical Examination: Requires the simultaneous availability of a doctor, a nurse, and a specific room type (A or B) depending on the patient's assigned priority.

Optimization Goals

The simulation acts as a decision support tool to find the optimal staffing balance. The core success metrics were strictly defined:

1. Ensure the average wait time for ambulance patients (from building entry to the start of treatment) does not exceed 15 minutes.
2. Ensure the average wait time for walk-in patients is kept under 30 minutes.