Scholarlink

<p>Natural disasters, such as earthquakes and fires, can severely impede rescue operations by creating blocked roads and preventing emergency agents from reaching affected areas in time. Existing path-planning solutions in the RCRS System rely mainly on conventional A* routing and nearest-obstacle clearance strategies, which often ignore varying disaster conditions and tend to produce curved or inefficient clearance paths. To address these limitations, we propose AA*-SOC, an integrated framework combining Adaptive A* graph routing and Smoothed Obstacle Clearance for police-force agents. The AA* approach introduces a condition-aware weighting scheme that incorporates road geometry, obstacle characteristics, and disaster severity into the heuristic, while the SOC strategy complements routing through direction-aware obstacle removal that minimises turning and clearance redundancy. Experiments conducted across four real-world regions demonstrate consistent improvements over Dijkstra, A*, random A*, and D* baselines using the official RCRS rescue-score metric. The AA*-SOC framework achieves measurable gains in rescue efficiency across different scenarios, accelerates route clearance, and reduces unnecessary detours. Ablation studies further validate the individual contributions of both the AA* approach and the SOC strategy. Overall, the results indicate that incorporating disaster-aware routing and directional clearance leads to more reliable and time-efficient rescue performance.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Adaptive graph search for multi-agent rescue operations in natural disaster response

  • 期刊论文

  • 发表日期：

    2026年6月3日

查看全文
  • Qihua Lyu,
  • Cao Xin,
  • Yongsheng Yu,
  • Shan Xue,
  • Jian Yang,
  • Zhenyu Yang,
  • Amin Beheshti,
  • Jia Wu

摘要

Natural disasters, such as earthquakes and fires, can severely impede rescue operations by creating blocked roads and preventing emergency agents from reaching affected areas in time. Existing path-planning solutions in the RCRS System rely mainly on conventional A* routing and nearest-obstacle clearance strategies, which often ignore varying disaster conditions and tend to produce curved or inefficient clearance paths. To address these limitations, we propose AA*-SOC, an integrated framework combining Adaptive A* graph routing and Smoothed Obstacle Clearance for police-force agents. The AA* approach introduces a condition-aware weighting scheme that incorporates road geometry, obstacle characteristics, and disaster severity into the heuristic, while the SOC strategy complements routing through direction-aware obstacle removal that minimises turning and clearance redundancy. Experiments conducted across four real-world regions demonstrate consistent improvements over Dijkstra, A*, random A*, and D* baselines using the official RCRS rescue-score metric. The AA*-SOC framework achieves measurable gains in rescue efficiency across different scenarios, accelerates route clearance, and reduces unnecessary detours. Ablation studies further validate the individual contributions of both the AA* approach and the SOC strategy. Overall, the results indicate that incorporating disaster-aware routing and directional clearance leads to more reliable and time-efficient rescue performance.