About the Infinite Windy Firebreak Location problem

TL;DR

This paper models wildfire containment as an optimization problem on infinite graphs, proposing methods to select firebreaks under budget constraints to prevent unbounded fire spread.

Contribution

It introduces the Infinite Windy Firebreak Location problem, proves its computational complexity, and relates it to the Min Cut problem for specific graph classes.

Findings

Infinite Windy Firebreak Location is coNP-complete in restricted cases.

Polynomial cases are identified and analyzed.

Problem reduces to Min Cut in certain graph classes.

Abstract

The severity of wildfires can be mitigated adopting preventive measures like the construction of firebreaks that are strips of land from which the vegetation is completely removed. In this paper, we model the problem of wildfire containment as an optimization problem on infinite graphs called Infinite Windy Firebreak Location. A land of unknown extension is modeled as an infinite undirected graph in which the vertices correspond to areas subject to fire and edges represent the propagation of fire from an area to another. The construction of a firebreak on the territory is modeled as the removal of edges in both directions between two vertices. The number of firebreaks that can be installed depends on budget constraints. We assume that fire ignites in a subset of vertices and propagates to the neighbours. The goal is to select a subset of edges to remove in order to contain the fire and avoid burning an infinite part of the graph. We prove that Infinite Windy Firebreak Location is coNP-complete in restricted cases and we address some polynomial cases. We show that Infinite Windy Firebreak Location polynomially reduces to Min Cut for certain classes of graphs like infinite grid graphs and polyomino-grids.