A Comparative Analysis of Task Allocation Strategies for Resource-Constrained Multi-Agent Drone Swarms in Wildfire Response
Huan Vo, Constantino Álvarez Casado, Abhishek Kumar, Satu Tamminen, Petri Ihantola, Miguel Bordallo López, Le Nguyen
University of Oulu University of Jyväskylä
阅读操作
确认中在文库中上传 PDF 后可生成中文音频讲解。
摘要与影响
The coordination of autonomous drone swarms for wildfire response presents a key problem in dynamic task and resource management. This paper presents a comparative analysis of coordination strategies for a resource-constrained multi-agent drone swarm. The evaluation is conducted within an agent-based simulation that includes a probabilistic Cellular Automata (CA) fire model with multi-directional wind and heterogeneous fuel loads. The drone agents operate under constraints, including a limited water payload or battery charge that requires periodic resupply. We implement and assess three strategies: a decentralized heuristic, a context-aware auction mechanism where bids reflect local fire intensity and predictive risk, and a reinforcement learning (RL) agent based on Proximal Policy Optimization (PPO). The RL agent integrates an attention mechanism to extract relevant environmental cues and is informed by short-term fire spread forecasts to support proactive decision-making. The simulation results provide a quantitative comparison of these strategies based on fire containment, task completion time, and resource use. Although the auction mechanism offers robust performance, the predictive RL agent learns more proactive containment behaviors. This analysis quantifies the trade-offs across coordination strategies and informs the design of intelligent, resource-aware systems for autonomous wildfire response. The code is available at: https://github.com/Arritmic/oulu-wildfire-drones.
逐年被引趋势
关键指标
同类平均 = 1
同领域 · 同年份 · 同类型
Google Scholar 与 OpenAlex 的被引统计范围不同,数值存在差异属正常。
AI 辅助阅读
依据:摘要
可就本文提问;依据不足时会说明。
学术脉络
学科主题
工程Distributed Control Multi-Agent Systems
UAV Applications and Optimization · Modular Robots and Swarm Intelligence