A stochastic programming standard response model for wildfire initial attack planning

作者：Ntaimo Lewis^*; Arrubla Julian A Gallego; Stripling Curt; Young Joshua; Spencer Thomas

来源：Canadian Journal of Forest Research, 2012, 42(6): 987-1001.

DOI：10.1139/X2012-032

Wildfires are responsible for several civilians deaths and millions of dollars in property losses every year, on average. Wildfire containment is the result of an effectively performed initial attack. We formulate a two-stage stochastic integer programming standard response model for initial attack. The model assumes a known standard response needed to contain a fire of given size. The goal of the model is to contain as many fires as possible while minimizing the fixed rental and travel costs and the expected future operational costs. We report on a study based on district TX12, which is one of the fire planning units of the Texas Forest Service (TFS) in East Texas, involving seven operations bases and up to 28 dozers. We use our methodology to position the dozers based on 30- and 60-min maximum travel time restrictions and evaluate each deployment in terms of the number of fire-fighting resources positioned at each operations base and the expected number of escaped (contained) fires. We also consider deploying additional resources to operations bases that need them the most. The deployments made by our methodology provide several insights and show that the original distribution of resources in TX12 at the time of this study was not optimal. For example, more dozers were initially located at operations bases in areas of low density of fires while fewer dozers were located at bases in areas of high density of fires.

• 出版日期2012-6