The two-way ten-lane expressway has the significant characteristics of “large traffic volume, mixed vehicle types, and heavy loads”, which makes the impact of traffic flow status on accident risk present nonlinear characteristics. Traffic flow fluctuations not only directly affect the probability of accidents, but also amplify the spatiotemporal differences in rescue needs through mechanisms such as lane occupancy time and accident chain reactions. Therefore, the essence of resource allocation on a two-way ten-lane expressway is the “spatiotemporal matching problem between dynamic risks and limited resources”, which requires both quantifying the spatiotemporal evolution of risks and coping with the high uncertainty of the traffic system.
Aiming at the problem of inefficiency of traditional empirical resource allocation under complex traffic conditions, this study proposes a dynamic optimization framework based on multidimensional risk assessment for emergency rescue resource allocation. In this framework, firstly, the entropy weight method and fuzzy comprehensive evaluation are combined to construct a risk quantification model using historical accident data and real-time traffic characteristics to achieve fine risk classification of road sections. Secondly, a multi-objective optimization model is established with the goal of minimizing risk-weighted costs and maximizing risk-weighted resource demand satisfaction, and considering constraints such as mandatory requirements for key equipment in high-risk areas and minimum site configuration. At the same time, the improved NSGA-II algorithm is used to effectively solve the contradiction between cost and utilization efficiency in emergency rescue resource allocation through adaptive non-dominated sorting, hybrid genetic operators and dynamic penalty mechanism.
Experimental results show that the improved NSGA-II algorithm is superior to the traditional method in terms of Pareto front distribution, convergence speed and actual resource allocation effect. Compared with the traditional scheme, the method proposed in this study reduces the resource allocation cost by 35.5%, increases the risk-weighted resource demand satisfaction rate by 1.9%, and expands the resource coverage of high-risk areas by 13.8%. This study provides scientific decision-making support for emergency response in complex road networks and offers a practical optimization approach for highly dynamic traffic emergencies.