Annals of Operations Research vol:219 issue:1 pages:433-456
A heuristic solution method to identify a predefined number of cells in raster maps optimizing both multiple intrinsic cell characteristics and cell patch compactness (Vanegas et al., 2008), is extended so that also spatial interaction is taken into account. Spatial interaction is e.g., present in the flow of sediment between cells in a rasterised watershed as modified by the presence or absence of tree cover in a given cell. The resulting HLC-method is elaborated and tested in the context of a site location problem for reforestation in which the following criteria are considered: 1) the sediment flow reaching the outlet of the watershed is minimized, 2) two intrinsic environmental performances of the cells are maximized/minimized, and 3) the selected cells form a compact patch. To evaluate the performance of the HLC, its results are compared to the ones obtained with an Integer Programming (IP) formulation. This comparison suggests that the HLC identifies high quality, near-to-optimal patches of cells and that its computational efficiency makes it applicable to regional sized cases.