Development of a Genetic Algorithm Model for a Multiple-Objective Forest Harvesting and Zonation Problem

Abstract

Most of the forested lands in North America are managed for multiple objectives and forest management plans are designed in a manner to obtain or maintain sustainable forest management certification. One strategy for achieving some of these goals is forest zonation, which allows for different intensities of forest management (including reserve areas) in different zones of the forest. The focus of this thesis is the development of a spatially explicit forest estate model which simultaneously allocates forest land to different management intensity zones and harvest periods with the goal of satisfying multiple management objectives. The model was initially developed using mixed integer goal programming. This helped to identify a basic model structure which was used to guide the development of a genetic algorithm-based implementation. The development of the model, particularly the setting of goal weights to describe the decision maker's preferences is described in detail.