Optimizing Resource Allocation for Urban Public Safety Risk Governance

Abstract

Rapid urbanization, compound hazards, and interagency dependencies increase the importance of timely and reliable emergency resource allocation. This study develops a risk-responsive robust multi-objective allocation framework (RRMA) connecting risk assessment, demand generation, network accessibility, and deployment decisions. A four-dimensional index represents hazard, exposure, vulnerability, and recovery capacity. Facility activation, resource pre-positioning, and scenario-dependent distribution are subject to budget, capacity, and minimum-service constraints, while robustness is expressed through the worst weighted loss over a finite scenario set. The scenario analysis covers six urban functional area types and compares baseline-risk, flooding, and compound-risk synthetic samples. Aggregate results show that, relative to the NSGA-II benchmark, RRMA increases high-risk demand coverage by 3.2–3.7 percentage points, reduces mean response time by 2.9–3.1 minutes, lowers unmet demand by 2.7–2.8 percentage points, and reduces the coverage Gini coefficient by 0.031–0.032. These differences indicate that risk-responsive allocation can reconcile service efficiency and spatial equity in the synthetic settings examined. The comparisons provide descriptive evidence at the framework level and do not identify the independent contribution of individual model constraints. The study provides an explicit basis for coordinating risk priorities, minimum-service protection, and cross-district resource allocation under limited budgets. 快速城市化、复合灾害以及部门间相互依赖，对应急资源配置的及时性与可靠性提出了更高要求。本研究构建了一套风险响应型鲁棒多目标资源配置框架（risk-responsive robust multi-objective allocation, RRMA），将风险评估、需求生成、网络可达性与资源部署决策纳入统一分析框架。采用灾害危险性、暴露度、脆弱性和恢复能力四个维度构建风险指数，并在预算、容量及最低服务水平约束下，对设施启用、资源预置和不同情景下的资源分配进行联合优化；同时，通过有限情景集下的最坏加权损失刻画配置方案的鲁棒性。情景分析涵盖六类城市功能区，并基于合成样本比较基准风险、洪涝灾害和复合风险三种情景。汇总结果显示，与 NSGA-II 基准方法相比，RRMA 将高风险需求覆盖率提高了 3.2–3.7%，平均响应时间缩短了 2.9–3.1 分钟，未满足需求比例降低了 2.7–2.8%，覆盖率基尼系数降低了 0.031–0.032。上述结果表明，在本文所设定的合成情景中，风险响应型资源配置能够兼顾服务效率与空间公平。相关比较提供的是框架层面的描述性证据，尚不能识别各项模型约束的独立贡献。本研究为有限预算条件下统筹风险优先级、最低服务保障与跨区域资源配置提供了一套明确的分析与决策框架。