Research on multi-objective scheduling optimization of emergency resources based on simulated annealing algorithm
Abstract
Emergency departments face persistent challenges including prolonged patient wait times, unbalanced resource utilization, and high staff overtime. To address these issues, this study proposes a multi‑objective resource scheduling model based on real hospital data and optimizes it using a simulated annealing (SA) algorithm. The model discretizes time slots and integrates multiple resources (doctors, consultation rooms, beds), with objectives of minimizing weighted waiting time and labor cost under triage‑priority and maximum‑wait constraints. A neighborhood operation set (patient exchange, cross‑doctor reallocation, shift fine‑tuning) was designed to fit clinical workflows. Simulation on annual 2023 data from a tertiary hospital showed that, compared to the current system, the SA‑optimized schedules reduced average waiting time by 20–35%, lowered the overall threshold-violation rate by 46.8% (26.7 to 14.2%), and balanced resource utilization across weekday, peak‑day, and night‑shift scenarios. These gains were not cost-neutral: total daily labour cost rose by 11.4% (CNY 31,680 to 35,280) and cost per patient visit by 10.6% (CNY 102.3 to 113.1), while overtime hours fell by 29.7% and overtime cost by 30.2%. The SA-based framework converged reliably and retained its ordering across scenarios, and is best described as a candidate decision-support component that buys substantial improvements in timeliness, resource equity and staff workload at a quantified increase in labour expenditure. Although the framework was validated only in a simulation environment calibrated with one year of real ED data and has not yet undergone prospective real-world deployment, the results suggest that it constitutes a potentially deployable decision-support component for intelligent ED scheduling. All results derive from a single-centre in-silico evaluation calibrated on one tertiary emergency department; prospective and multi-centre validation is required before any claim of generalisability or operational readiness can be supported.
Related
- Multi-objective casualty distribution and real-time coordinated transportation scheduling of emergency medical materials
- Optimization of multi-objective flexible scheduling for container rail mounted gantry cranes in railway logistics yards
- Improved genetic algorithm based intelligent scheduling model and efficient solution method for green financial resources
- Multi-objective energy scheduling in multiple energy system considering economic and environmental objective functions
- Coupling calibrated aerial perception to adaptive UAV emergency logistics scheduling: a YOLOv8–reinforcement learning closed-loop proof of concept
BibTeX
@article{song2026research,
title = {{Research on multi-objective scheduling optimization of emergency resources based on simulated annealing algorithm}},
author = {Zimo Song},
journal = {Scientific Reports},
year = {2026},
month = oct,
doi = {10.1038/s41598-026-72989-8},
url = {https://www.nature.com/articles/s41598-026-72989-8},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Research on multi-objective scheduling optimization of emergency resources based on simulated annealing algorithm
AU - Zimo Song
PY - 2026
DA - 2026/10/01
JO - Scientific Reports
DO - 10.1038/s41598-026-72989-8
UR - https://www.nature.com/articles/s41598-026-72989-8
AB - Emergency departments face persistent challenges including prolonged patient wait times, unbalanced resource utilization, and high staff overtime. To address these issues, this study proposes a multi‑objective resource scheduling model based on real hospital data and optimizes it using a simulated annealing (SA) algorithm. The model discretizes time slots and integrates multiple resources (doctors, consultation rooms, beds), with objectives of minimizing weighted waiting time and labor cost under triage‑priority and maximum‑wait constraints. A neighborhood operation set (patient exchange, cross‑doctor reallocation, shift fine‑tuning) was designed to fit clinical workflows. Simulation on annual 2023 data from a tertiary hospital showed that, compared to the current system, the SA‑optimized schedules reduced average waiting time by 20–35%, lowered the overall threshold-violation rate by 46.8% (26.7 to 14.2%), and balanced resource utilization across weekday, peak‑day, and night‑shift scenarios. These gains were not cost-neutral: total daily labour cost rose by 11.4% (CNY 31,680 to 35,280) and cost per patient visit by 10.6% (CNY 102.3 to 113.1), while overtime hours fell by 29.7% and overtime cost by 30.2%. The SA-based framework converged reliably and retained its ordering across scenarios, and is best described as a candidate decision-support component that buys substantial improvements in timeliness, resource equity and staff workload at a quantified increase in labour expenditure. Although the framework was validated only in a simulation environment calibrated with one year of real ED data and has not yet undergone prospective real-world deployment, the results suggest that it constitutes a potentially deployable decision-support component for intelligent ED scheduling. All results derive from a single-centre in-silico evaluation calibrated on one tertiary emergency department; prospective and multi-centre validation is required before any claim of generalisability or operational readiness can be supported.
SN - 2045-2322
ER -