Article ID: | iaor20052202 |
Country: | Netherlands |
Volume: | 157 |
Issue: | 2 |
Start Page Number: | 409 |
End Page Number: | 428 |
Publication Date: | Sep 2004 |
Journal: | European Journal of Operational Research |
Authors: | Tangian Andranik |
Keywords: | allocation: resources, artificial intelligence: decision support, programming: quadratic |
University budgets are redistributed proportionally to five target criteria. The approach has been implemented in several models and applied to 15 North-Rhine Westphalia state universities. In the given paper, the problem is considered from the viewpoint of the general techniques for constructing objective functions for evaluating alternatives. It enables us to formulate three additional models aimed at best respecting the status quo: with minimizing absolute changes of actual budgets, with minimizing relative changes of actual budgets, and with minimizing changes of individual rules for budget accounting. The models are based on either linear, or quadratic objective functions, the latter being considered under several optional assumptions: local monotonicity, global monotonicity, concavity, and convexity. The paper is illustrated with computer results.