Abstract:
Mathematical models have the potential to provide a cost-effective, objective,
and flexible approach to assessing management decisions, particularly when
these decisions are strategic alternatives. In some instances, mathematical
model is the only means available for evaluating and testing alternatives.
However, in order for this potential to be realized, models must be valid for
the application and must provide results that are credible and reliable. The
process of ensuring validity, credibility, and reliability typically consists of
three elements: verification, validation, and calibration.
Model verification, validation and calibration are essential tasks for the
development of the models that can be used to make predictions with
quantified confidence. Quantifying the confidence and predictive accuracy of
model provides the decision-maker with the information necessary for making
high-consequence decisions.
There appears to be little uniformity in the definition of each of these three
process elements. There also appears to be a lack of consensus among model
developers and model users, regarding the actions required to carry out each
process element and the division of responsibilities between the two groups.
This paper attempts to provide mathematical model developers and users
with a framework for verification, validation and calibration of these models.
Furthermore, each process element is clearly defined as is the role of model
developers and model users.
In view of the increasingly important role that models play in the
evaluation of alternatives, and in view of the significant levels of effort
required to conduct these evaluations, it is important that a systematic
procedure for the verification, validation and calibration of mathematical
models be clearly defined and understood by both model developers and
model users.
Machine summary:
J. Humanities (2012) Vol. 19 (4): (15-32) A Note on Models' Verification, Validation and Calibration Mohammad Ali Khatami Firouzabadi1 Received: 2011/10/4 Accepted: 2012/4/14 Abstract Mathematical models have the potential to provide a cost-effective, objective, and flexible approach to assessing management decisions, particularly when these decisions are strategic alternatives.
This paper attempts to provide mathematical model developers and users with a framework for verification, validation and calibration of these models.
Definition of the Verification, Validation and Calibration Typically, developers create models and perform initial model evaluations to provide model users with a level of assurance that the model is reliable and realistic.
Unfortunately, it is difficult to determine the success or failure of the calibration process by strictly examining the field data and the selected input parameter values because input parameters are incompletely known and there may be unknown effects on the model.
Model users are often not able to A note on models' verification, validation and calibration determine the impact that input parameter values have on the selected model.
If the results are within the level of accuracy specified by the developer, the verification process is A note on models' verification, validation and calibration considered successful.
A note on models' verification, validation and calibration 3- The generation of model results first require that input parameter values be calibrated using the collected field data.
Even if the discrepancies between the model results and the field data are very small, it is possible that compensating errors in the selection of the input parameter values may mask fundamental flaws in the model logic.