A NEW HEURISTIC TO SOLVE SPATIALLY CONSTRAINED LONG-TERM HARVEST SCHEDULING PROBLEMS

Because of capability limitations of the integer programming solution technique, a new heuristic algorithm was developed to solve spatially constrained long-term harvest scheduling problems. The proposed algori thm can handle multiple harvesting for each harvest unit over a long t ime horizon. The heuristic utilizes random ordering heuristic optimiza tion and the PATH algorithm adapted from stand level optimization. Emp loying the proposed algorithm, a harvest scheduling system was constru cted. The performance of the proposed algorithm is presented compared to the branch-and-bound algorithm in terms of the computational time a s well as the objective value. Using two example forests, solutions by the proposed algorithm are stable in terms of the objective value and have harvest flow fluctuation much less than 3%. For short-term probl ems, solutions by the proposed algorithm tend to be optimal. For those problems, for which an optimal solution is found by the branch-and-bo und algorithm, the solution can produce an objective value with deviat ion less than 2% from the optimum. The proposed algorithm yields bette r solutions for long-term problems than the branch-and-bound algorithm with the 1,000,000 limited number of iterations, and the lower bound derived by the proposed algorithm. Computational results reveal that a s the time horizon increases, the proposed algorithm significantly and increasingly outperforms the ''limited'' branch-and-bound algorithm i n terms of required computational time. The advantage of the proposed algorithm results from partitioning the problem into subproblems perio d by period using the PATH algorithm, and defining the objective funct ion of the subproblem by minimizing absolute infeasibility on harvest flow constraints at each period under a two-period sequential feasibil ity condition.