A review of some of the current thinking pertaining to suboptimization and decentralization is conducted. Rueflis generalized goal decomposition model is discussed with emphasis placed upon description of the informational requirements necessary for the iterative process and the constraints necessar for the model to function. The generalized goal decomposition model is extended to include quadratic deviations from goals. The nature of weighting factors in an organizational context is examined. Some current methods for determining weighting factors for goal programming problems are presented. A goal programming method using a separable objective function for the management level is proposed for determining weighting factors.