TY - JOUR
T1 - Scheduling forest core area production using mixed integer programming
AU - Wei, Yu
AU - Hoganson, Howard M.
PY - 2007/10/1
Y1 - 2007/10/1
N2 - Core area, the area of mature forest protected by a buffer area from edge effects of surrounding habitats, is an important spatial measure describing forest ecological conditions. Three alternative mixed integer programming (MIP) formulations are presented for explicitly scheduling core area production in a forest management scheduling model. Formulations utilize detailed data preprocessing that develops a set of influence zones. Each influence zone identifies an area of the forest that can produce core area. Each zone is influenced by a unique combination of management units (stands) of the forest. The assumed width of the buffer surrounding core area affects both the number of zones in the forest and the number of stands associated with each zone. Numerous test cases were applied, varying the MIP formulation used to describe core area production, the assumed buffer for core area (50 or 100 m), and the set of additional forest-wide constraints to control harvest levels and core area production levels over time. Solution times varied substantially between the alternative MIP formulations. Solution times were substantially less for the formulation that used more, but simpler, spatial constraints. Solution times for large test cases suggest that real-world applications are likely feasible.
AB - Core area, the area of mature forest protected by a buffer area from edge effects of surrounding habitats, is an important spatial measure describing forest ecological conditions. Three alternative mixed integer programming (MIP) formulations are presented for explicitly scheduling core area production in a forest management scheduling model. Formulations utilize detailed data preprocessing that develops a set of influence zones. Each influence zone identifies an area of the forest that can produce core area. Each zone is influenced by a unique combination of management units (stands) of the forest. The assumed width of the buffer surrounding core area affects both the number of zones in the forest and the number of stands associated with each zone. Numerous test cases were applied, varying the MIP formulation used to describe core area production, the assumed buffer for core area (50 or 100 m), and the set of additional forest-wide constraints to control harvest levels and core area production levels over time. Solution times varied substantially between the alternative MIP formulations. Solution times were substantially less for the formulation that used more, but simpler, spatial constraints. Solution times for large test cases suggest that real-world applications are likely feasible.
UR - http://www.scopus.com/inward/record.url?scp=38149118672&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=38149118672&partnerID=8YFLogxK
U2 - 10.1139/X07-033
DO - 10.1139/X07-033
M3 - Article
AN - SCOPUS:38149118672
SN - 0045-5067
VL - 37
SP - 1924
EP - 1932
JO - Canadian Journal of Forest Research
JF - Canadian Journal of Forest Research
IS - 10
ER -