|
Brief description: Realistic instances of wildfire spread in landscapes of different sizes and features, modeled as a network. The spread travel times between adjacent nodes have been calculated according to the methodology from Belval et al. (2015) and Wei et al. (2011). This methodology takes fire behaviour parameters such as maximum spread rate, maximum spread direction and the shape of the fire ellipse to calculate spread travel times. |
||
|
Mission: To stop the spread minimizing the value of the are burned using two teams |
||
|
Network: Each landscape is modeled as a network, which nodes are the centroids of each raster cell. Two nodes are joined by an arc if they represent adjacent patches in the landscape. Instances range from a 5×5 grid to a 7×7 grid. 25 instances are created for each landscape size. A larger Case Study of size 10×10 is also created. Two types of problems are considered, when the network is a complete graph for the teams (every node is reachable from any other node) called the TOPVTW─nodes. And when the teams can only move between adjacent nodes called the TOPVTW─paths |
||
|
Data: |
||
|
Data sources: The fire behaviour parameters have been calculated using FlamMap, a fire simulator which considers factors such as the slope, the wind or the type of vegetation. The target landscapes are terrain patches created from the available raster information provided by the LandFire program. Each raster cell size is 30x30m big. Weather conditions are assumed to be constant; wind direction and intensity have been set to 45 degrees azimut and 30km/h respectively. The teams are assumed to be homogeneous, having all common features, and their travel times are calculated based on their velocity and the slope and aspect of the terrain provided by the LandFire program. |
||
|
Tested in: The cases are tested in (1), which introduces a new variant of the Orienteering Problem called “The Team Orienteering Problem with Variable Time Windows”
|
