Road-network location heuristics for the tactical harvest-scheduling model

Abstract

In tactical planning in hierarchical forest management, cut-blocks are selected for maximizing revenue and road networks are allocated at minimal cost in order to maximize profit. The selected cut-block set and requisite road-network, connecting the cut-blocks, therefore have an interdependent relationship. The location of these two elements in tactical planning must therefore be considered simultaneously in a tactical harvest-scheduling model. This integration presents a major computational challenge, especially with regard to the execution time required to find an optimal solution to the tactical harvest-scheduling model. The objective of this thesis isto explore the influence of different road location heuristics, used within the tactical harvest scheduling model, upon the model’s execution time and solution quality. We nested the three different types of roadlocation heuristics within the harvest-scheduling model in order to evaluate their effectiveness by three criteria: execution time, road construction cost and objective function value. In addition, after the tactical model was run, we executed and evaluated the usefulness of a road network repair algorithm, designed to improve further the solution of the road-network location generated by the tactical harvest-scheduling model. The thre heuristics were evaluated on a real-world dataset, representing a section of the Kenogami forest in Ontario, Canada. Our result show: i) that the Shortest Path Origin Heuristic (SPOH) achieved the fastest execution time and lowest construction cost when integrated within the tactical harvest-scheduling model; and ii) that the road network repair algorithm successfully lowered the road network costs and thereby increased the objective function value of all solutions generated using the tactical planning model. These results are significant for two reasons: first, they show that the choice of the road network heuristic used within a tactical planning model can have a major influence on the model’s solution quality; and second, that the use of a road repair algorithm, on the solution generated using a tactical model, is of major economic value in forest management planning.