| Abstract: | In this paper we consider a classical model of price-directive decision making in hierarchical organizations, namely Dantzig-Wolfe decomposition, well known from single-objective programming. Here, however, we allow for preference differences that are usually observed between a central unit and a subunit in an organizational structure. The procedure models, therefore, how decision making may evolve in a context with decentralized information as well as intra- and interpersonal conflicts. Several numerical experiments with the procedure have been performed. The results of those experiments demonstrate that the procedure in many instances converges towards an efficient solution, despite the differences in the criteria weights used in the master problem and the subproblem respectively. This suggests that some amount of goal discordance need not prohibit efficient decision making in a hierarchical organization. |
