Optimization of pre-hospital emergency facility layout in Nanjing: A spatiotemporal analysis using multi-Source big data

Abstract

Amid the escalating conflicts between urban demography, resource availability, and environmental constraints, there is an accelerating demand for emergency medical services prompted by a spectrum of factors including diseases, natural calamities, and unforeseen events. This growth is further accentuated by an imbalance in the allocation of emergency facilities, amplifying public anxiety and underscoring the urgent necessity for scientifically informed optimization of pre-hospital emergency facilities. To this end, the present study articulates specific optimization goals for facility deployment and introduces a nuanced set coverage optimization model that integrates spatiotemporal determinants of emergency service requirements. The model is fortified with a constellation of constraints, including spatial constraint, temporal constraint, and coverage constraint. For spatial constraints, we use the Monte Carlo simulation to predict the spatial distribution of emergency demands. Temporal constraints involve determining the actual travel time matrix from predicted demand points to candidate sites. Coverage constraints specify the effective demand coverage rate. This study uses Nanjing as a case example, utilizing multisource big data, including ambulance GPS logs from June 2016 to May 2017, Amap traffic congestion indicators, and survey data on existing emergency facilities in Nanjing City. By preprocessing and analyzing the existing data, this study thoroughly investigates the spatiotemporal distribution of emergency demands and the impact of traffic congestion on emergency service effectiveness. Consequently, the model incorporates constraints that ensure, under the specified planning and actual traffic conditions, 95 % of the simulated emergency demands can be met within an 8-minute on-route time. Locations for pre-hospital emergency stations are optimized using a genetic algorithm, achieving the best solution at the 120th iteration. Verification confirms 134 optimal sites; after excluding 52 existing sites, 82 potential sites are identified. The new layout plan has reduced Nanjing’s average emergency response time from 18.6 min to 12 min. Additionally, under peak and average traffic conditions, emergency demand coverage rates improved from 76.92 % and 83.18 % at 15 min to 95.61 % and 98.10 % at 12 min, respectively. These results demonstrate that the new layout significantly enhances practical application effectiveness. The approach presented in this paper addresses the previously overlooked randomness of emergency incidents and traffic conditions, offering innovative strategies for improving the efficiency and effectiveness of emergency station planning and site selection models.

Keywords

• Emergency medical services
• Spatiotemporal optimization
• Multi-modal big data
• Genetic algorithm
• Constraint modeling
• Nanjing city