Howard M. Hoganson

HEURISTICS IN MULTI-OBJECTIVE FOREST MANAGEMENT

Heuristics have been used extensively to support forest management scheduling in the last two decades. The need for spatial definition, and the combined shortcomings of available technology and traditional mathematical programming approaches, sparked interest in alternative forest management scheduling techniques in the early 1980s. Concerns with the environmental impacts of forest management options further encouraged the development of heuristics to address adjacency relationships in harvesting decisions. More recently, heuristics have been used to target other multi-objective management issues. Namely, they have been used to provide information to help sustain both traditional forest products flows (e.g. timber and cork) and landscape structural characteristics (e.g., mosaic elements such as patch number and size, amounts of edge or interior space). In this chapter, we describe the current state of the art of heuristic application in forest management scheduling. Heuristic approaches are presented and discussed in the framework of forest management scheduling needs. Results from some heuristic research efforts are used to outline the application potential and shortcomings of these techniques.

Coordinating Management Decisions of Neighboring Stands with Dynamic Programming

Stands are typically thought of as the basic analysis area (AA) or building block of a forest. Historically, forest management planning has focused on how stand management activities can be coordinated across the forest to provide a sustainable flow of products over time. With increasing concerns about sustaining environmental conditions, spatial facets of the management situation also become a concern. Often, how a stand is managed impacts not only the condition of the stand, but also the condition of a “neighborhood” surrounding the stand. These impacts can have a lasting impact over time, thus adding a temporal dimension to spatial concerns. Spatial concerns add substantial complexity for analyzing management options because forests include many stands, each with its own potential unique neighborhood. Generally, it is too simplified to separate the forest into independent neighborhoods, ignoring conditions along neighborhood boundaries. In recent years adjacency constraints have received considerable attention as one policy tool for addressing spatial management objectives. Adjacency constraints set maximum limits on the size of harvest blocks. Specific management limitations for adjacency constraints vary, without any one accepted best set of rules. Typically, the rules involve only regeneration harvests associated with even-aged management. Rules may limit harvesting of adjacent stands within a given time period or set a maximum limit on the size (total area) of any contiguous harvest block. Rules involve an “exclusion period” during which harvesting is prohibited for adjacent stands around a harvest block. Definitions of exclusion periods also vary by user. Some assume a fixed time length for the exclusion period (e.g. 10 years) or tie its definition to the condition of the harvest block (e.g. not until the trees regenerated in the harvest block reach 20% of their height at rotation). Methods for incorporating adjacency constraints into forest planning models have received considerable

Techniques for Addressing Spatial Detail in Forest Planning

Specialized model solution approaches can be designed for forest management scheduling problems by utilizing an understanding of the problem. A specialized decomposition approach has made it possible to address larger problems. It has proven successful in applications. Concepts of moving windows from geographic information systems can be combined with dynamic programming (DP) techniques to address adjacency considerations in large problems. This DP approach and the specialized decomposition approach can likely be combined to help identify ways of sustaining both timber production and forest-wide spatial conditions such as the amount of forest edge or interior space. The specialized decomposition approach has been expanded to address spatial interactions between timber markets. Similar expansions seem plausible to address broader spatial environmental concerns related to forest biodiversity.

Tailoring a decomposition method to a large forest management scheduling problem in northern Ontario

AbstractA forest planning problem with multiple market locations and multiple products is formulated for a 3,166,000 ha forested region in northern Ontario. The resulting linear programming model is extremely large, with millions of decision variables and tens of thousands of constraints. A modelling method proposed by Hoganson and Rose (1984) was extended and applied successfully for eight model formulations. Optimal, near-feasible solutions were consistently produced in 200 to 300 iterations. Results showed definite, interpretable, patterns in the values of key dual variables that correspond to harvest level constraints in the primal LP and that tie the forest planning problem together.

Constrained Optimization for Addressing Forest-Wide Timber Production

Optimization models are important tools for integrating objectives in forest management planning. Timber management objectives are difficult to address without considering complex environmental conditions, other ownership objectives, and forest management policies. Typically, to learn about management opportunities, model applications involve multiple scenarios and tradeoff analysis. Study areas tend to be large with analyses challenging because of model size needs for addressing many facets of the situation. The trend is to decompose large problems into linked subproblems with feedback between analyses. A better understanding of forest management situations can be obtained by integrating forest management decision-making with timber supply chain analyses developed from wood users’ perspectives. Uncertainty surrounds forest management decision-making situations, with analyses expanding recently to address strategies for collecting forest inventory information and recourse opportunities to help reduce risks associated with an uncertain future.

Recognizing Spatial Considerations in Forest Management Planning

Purpose of ReviewThe aim is to help identify how spatial facets of forest management can be analyzed and better understood in strategic forest management planning. Focus is on stand-level spatial interdependencies potentially related to a wide range of considerations including wildlife habitat, invasive species, forest management regulations, and cost of harvest operations. Spatial facets addressed include adjacency and harvest area limitations, habitat connectivity, edge impacts, proximity considerations, and management options for restructuring stand shapes and sizes. Emphasis is on recent studies with direct connections to both forest management planning and problem structures of operations research.Recent FindingsModels related to explicit spatial facets of forest management are increasing in number, size, and complexity. Specialized approaches have been developed that are tailored to spatial facets of forestry problems. Improvements have also been made in ways of solving existing spatial models. Uncertainty is also being addressed in applications, and most recent studies tend to address multiple forest objectives.SummarySpatial interrelationships between stands are important considerations in forest planning. Operation research models can help explore the complex combinatorial nature of the situation. The need to better integrate multiple objectives over large landscapes is commonly suggested. Tradeoff analyses are important for decision makers to better understand forest-wide opportunities. New technology including parallel processing will help increase the practicality of model applications.