Abstract for: Solving optimization problems based on System Dynamics models

In the paper it is shown how optimization problems could be defined and solved using in the System Dynamics methodology. The presented approach is illustrated by three examples where effectiveness depends on made decisions. One of the widely recognized tool for verification of the decision rules in such environments is a System Dynamics approach which in general is based on differential–algebraic equations (DAE’s). SD models can be simply transformed to DAE’s form and computed with the help of advanced numerical procedures. As a result individual gets trajectories of variables which are de facto predictions of the environment behavior. While seeking an optimal decision, researcher needs to change values of particular parameters and compute new trajectories each time after such change. Mostly finding the optimal solution is time-consuming task. Other important issue of such approach is that reasercher never knows how far is he from the optimal decision rule. Therefore, having a dynamical model described by DAE’s, using optimal control theory with particular algorithms implemented in dedicated software we propose to define and solve optimal control problems. Its solution will give optimal trajectories of the variables of interest which could be verified against decision rules previously implemented in the model.