MODELIZATION METHODOLOGY (Systemic)
| Collection | International Encyclopedia of Systems and Cybernetics |
|---|---|
| Year | 2004 |
| Vol. (num.) | 2(2) |
| ID | ◀ 2176 ▶ |
| Object type | General information, Methodology or model |
A. CASELLES MONCHO and J.L. USO DOMENECH propose the following general systemic modelization methodology:
- “1. Identification of the elements to be considered according to the objectives of the modeler. Degree of detail or aggregation.
- “2. To give a name, a unit of measure and a range of possible values to each element determined. So the elements are converted in mathematical variables.
- “3. Identification of the causal dependencies among the variables considered.
- “4. To assign an equation or table of values to each dependency previously identified.
- “5. To program the model for a computer.
- “6. To verify that the model fits real data.
- “7. Validation of the model for another set of data obtained from another similar system.
- “8. To use the model for testing user policies” (1993, p.457).
The authors aim is the study of decision-making problems in enterprise management. This type of modelization could however be also used in other fields, as for example, ecology.
Their proposals should possibly be somewhat observed. In complex situations, it would be better to begin with a consensual critical scrutiny of the various participants objectives, in order to avoid man-made disasters due to unilateral views of some autocratic decision makers.
As to the 2nd point, it is not sure that all elements can be converted into mathematical variables, specially the motivations of the participants.
Causal dependencies may well be variable in time, i.e. at short, medium or long term.
Computer modeling (as stated by P. DENNING — see above), has some limits.
Finally, the model should be frequently recycled: Napoleon once said that the best way to lose a battle is to try to repeat a former victorious one.