System analysis techniques applied to the forrester world model—part I: Sensitivity analysis

Abstract

System analysis techniques can be used to study models such as the nonlinear dynamic feedback World Dynamics Model developed by Jay W. Forrester. Through application of sensitivity analysis it was discovered that the capital-investment parameters of the Forrester model exert the greatest influence on model behavior. Due to high sensitivity of these variables, careful inspection of their function and continual updating of their interrelations with other parameters must be performed to insure model accuracy. In many cases perturbation of a multiplier produces the same relative effect as an equal percentage change in one of the parameters designed as control variables. A change in the relation of the multiplier and its dependent variable offers an alternative to the direct control policies suggested by Forrester.