AbstractA network flow model is used for budget allocation in a multi-campus institution over several years. The network had a hierarchical structure with three hierarchies. The network's lower bounds represent the minimal requirements for each component. The upper bounds represent the predicted future requirement. They are predicted via a cost simulation model where several volume indices, policy variables and some environmental factors are used. Thus the overall planning process combines simulation with optimization.The model maximizes the additive utility function of the institution, expressed as a weighted summation over the preferences of the three hierarchies. The preferences are evaluated by using a multi-goal approach, utilizing pairwise comparisons and the eigenvalue prioritization technique developed by T. Saaty.