Tuesday, August 25, 2015

Subjectively biased objective functions

Le MENESTREL Marc, VAN WASSENHOVE Luk
Access the publisher's website
EURO Journal on Decision Processes (forthcoming)

The maximization of an objective function is a cornerstone of OR/MS modeling. How can we integrate subjective values within these models without weakening their scientific objectivity? This paper proposes a methodological answer that maintains the objective function and relaxes the maximization principle. We introduce a class of biased models that combine an objective function with a “subjective” factor that biases the maximization of such a function. We present the main properties of these models as well as the axiomatic foundations that allow for the rigorous measurement of biasing factors. We invite OR/MS scholars to participate in the development of practical applications integrating ethical and sustainability values.