Abstract :
Simulation is a very powerful technique, but it is essential that that power is harnessed and controlled. It is all too easy to become fascinated by the technique itself and lose sight of the original goals that were being sought. The author sets out an approach to model building which provides an effective method of control on such studies. The method requires specification of goals first and the reasons for this are explained. Once this has been done and the decision to proceed has been taken, the temptation to begin writing the detail of the model immediately must be avoided. As a result of experience in developing models, an approach has been devised which concentrates initially on defining a data structure that the model will be built on, and on producing an outline model. As the project progresses, there should then be regular progress meetings to review the status of the project. These should ensure that the original objectives are still being met. If the end user of the model is a different individual than the developer, then early versions of the model should also be left with the end user