In this paper we address the problem of routing the firefighting resources available for an extended attack to a wildfire. We conceive a mixed integer programming (MIP) model that integrates fire spread and resource operations (movement, direct, and indirect attacks), both spatially and temporally. We consider lexicographic objectives related to fire spread—the protection of sensible points and the minimization of the burned area—and to resources—minimizing the cost of their usage and the duration of their routes. At the core of the model is a set of networks: one for fire and one for each type of resource, linked by covering relations representing suppression and safety. This work extends existing MIP models for fire spread and resources positioning (in particular, extending Alvelos et al. [6]). We report on some preliminary computational experiments to validate the model and assess how difficult it is to solve instances derived from an actual landscape with fire tranmission times obtained by the Rothermel’s fire behaviour model.

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

A MIP Model for Wildfire Extended Attack

  • Filipe Pereira e Alvelos,
  • Marco Marto,
  • André Bergsten Mendes

摘要

In this paper we address the problem of routing the firefighting resources available for an extended attack to a wildfire. We conceive a mixed integer programming (MIP) model that integrates fire spread and resource operations (movement, direct, and indirect attacks), both spatially and temporally. We consider lexicographic objectives related to fire spread—the protection of sensible points and the minimization of the burned area—and to resources—minimizing the cost of their usage and the duration of their routes. At the core of the model is a set of networks: one for fire and one for each type of resource, linked by covering relations representing suppression and safety. This work extends existing MIP models for fire spread and resources positioning (in particular, extending Alvelos et al. [6]). We report on some preliminary computational experiments to validate the model and assess how difficult it is to solve instances derived from an actual landscape with fire tranmission times obtained by the Rothermel’s fire behaviour model.