Skip to main content
Open Access Publications from the University of California

Restoring Transparency to Computational Solutions

  • Author(s): Geoffrion, Arthur
  • et al.

Computational methods in support of decision-making have grown greatly in power during recent decades owing in large measure to Moore's Law. Nothing like this law operates in the realm of mathematical models, which has led to an increasingly lopsided “mind share” in favor of computation at the expense of mathematics among decision support system developers. The increased emphasis on computational methods is a mixed blessing, for these seldom excel at revealing why the solutions they yield are what they are. Yet in many situations, decision makers and policy makers need to understand the why behind these solutions in order to convince themselves and others of the need for action or to deepen their own understanding of the system under study. This paper advocates, with two detailed illustrations, conceptually simple models, arguments, and spreadsheets as adjuncts to complex computational models to help explain important aggregate properties of detailed computational solutions. This can improve the transparency, and hence the value, of such solutions.

Main Content
Current View