Abstract
 Linking LP Models with Spatially-Explicit Models for Forest Planning Track: Modeling Author(s): Michael Buell Aspatial linear programming (LP) models of forest resource development are polygon-based, driven by a management objective, produce an "optimum" schedule, quick and inexpensive to solve, and are extensible to financial and log allocation analyses. However, these aspatial models cannot adequately represent the spatial relationships between critical features on the ground or patterns of disturbance. Raster and polygon-based spatial models can explicitly represent these critical spatial relationships, but are typically driven by heuristics (e.g., extend a road network) and are unlikely to be optimal. Furthermore, spatially-explicit approaches are much less efficient than the aspatial models. We have developed a methodology that links an aspatial LP-based model with a spatially-explict sub-model, and preserves the advantages of both approaches. Reduced-cost coefficients from the LP model guide the spatial location of the disturbance schedule, ensuring that the spatial solution remains optimal. It is expected that this approach can be used outside the forestry realm. Michael Buell Cortex Consultants Inc. 2a-1218 Langley Street Victoria , British Columbia V8W 1W2 CA Phone: (250) 360-1492 Fax: (250) 360-1493 E-mail: mbuell@cortex.ca |
|||||
