Revisiting the issue of model validation in OR: An epistemological view

Abstract

Debates about model validation in Operational Research (OR) are both numerous and diversified. They all imply, in one way or another, the more general question of knowledge production. The problem of model validation in OR then unavoidably holds an epistemological dimension, and it is this dimension that will be privileged here. It will be seen that three successive and rival epistemological perspectives on the production of scientific knowledge have dominated the 20th century. They are the philosophical, the historical and the sociological perspectives. Each of them brings forward an image or representation of science which sheds a particular light on the issue of model validation in OR. Moreover, it will be seen that the generally prevalent view of models and of their validation in OR can be traced back to the philosophical perspective, a perspective which has been, for many decades, widely questioned in the field of epistemology. This leads to some conclusions on the process of knowledge production in the field of OR and, more particularly, on model validation. Overall, the paper aims at demonstrating the usefulness of epistemology for understanding how knowledge is produced and validated in the field of OR.