Dispatching, Positioning and Routing Resources for Wildfire Initial Attack

• Published in: Networks Vol. 86; no. 1; pp. 80 - 102
• Main Authors: Alvelos, Filipe, Marto, Marco, Mendes, André
• Format: Journal Article
• Language: English
• Published: Hoboken, USA John Wiley & Sons, Inc 01.07.2025; Wiley Subscription Services, Inc
• Subjects: Forest fires; Helicopters; Integer programming; Land use; Linear programming; Mixed integer; mixed integer programming; network optimization; Optimization; Roads & highways; Shortest-path problems; Travel time; wildfire suppression; Wildfires; Wind direction
• ISSN: 0028-3045; 1097-0037
• DOI: 10.1002/net.22278

ABSTRACT Wildfire is a global issue that requires contributions from different fields to address its multiple and potentially severe impacts. The subject of this paper is wildfire suppression. In particular, a mixed integer programming (MIP) model is formulated for dispatching, positioning, and routing firefighting resources (e.g., crews and helicopters) in the initial attack. Following the minimum travel time principle, wildfire spread is modeled as the shortest path in a network, as is common in the literature for this type of approach. As for modeling resource movements and attack positions, it is proposed that each type of resource has its own network (e.g., roads) and interaction with the fire spread network. A heuristic, which alternates between MIP (for dispatching and positioning) and shortest path algorithms (for routing) is also conceived, and shown to provide significantly better solutions than solving the complete MIP model. Experiments with data from an actual landscape in Portugal were conducted with pyO3F (a Python framework being developed under the scope of the project “An optimization framework for reducing forest fire”). pyO3F builds the networks of fire and resources from vector and raster files (e.g., land use, altitudes, roads) and parameters from the user (e.g., resolution, fire behavior model). After an optimization approach is applied to a given scenario (characterized by a set of ignition nodes and a wind direction and intensity), pyO3F returns results in text (e.g., the burned area in given instants) and vector files (e.g., fire spread, resources movements, and attacks).