2 METHODS FOR SOLVING STAND MANAGEMENT PROBLEMS BASED ON A SINGLE TREE MODEL

Two methods are presented for solving the stand management problem where the growth dynamics are depicted with a single tree model. With the nongradient method the problem is recast into a combinatorial problem, which in turn is solved with the method of simulated annealing. The gradient method, of which three versions are given, utilizes a combination of nonlinear and mixed integer techniques for solving the problem. In both cases, a linear programming problem solves the single tree harvest selection problem. In tests with nine sample problems, all methods are within 1% of the best solution found. The computational effort of the nongradient method is about one order of magnitude greater than that of the gradient methods. The results indicate the advantage of being able to divide the problem into a suitable hierarchy, since management problems based on single tree models are often too complex to be solved in one step. At each level an appropriate technique can then be applied.