A Hybrid Heuristic Algorithm for Forest Harvest Decision

In this article, a hybrid heuristic algorithm based on Genetic Algorithm and Hooke and Jeeves is described for solving a complicated forest harvest decision problem, which involves optimization of thinning and final felling under price risk for a mixed species stand of spruce and pine. The strategy consists of two optimal stock level functions and one reservation price function; in which, there are ten variables that need to be optimized. The hybrid heuristic algorithm consists of two stages. At the first stage, a Genetic Algorithm is applied to generate initial candidate solutions. At the second stage, the Hooke and Jeeves is applied to find the optimal solutions using these initial solutions. As a benchmark, a pure Genetic algorithm, Hooke and Jeeves, and Powell search are also tested. Results show that the hybrid heuristic algorithm is the best one among all of the tested algorithms. The Genetic Algorithm ranks second, the Hooke and Jeeves the third and the Powell search is the worst.