COMPLEXITY IN DYNAMIC-SYSTEMS

In research dealing with complex problems it is common practice to use computer-simulated dynamic systems. The complexity of such systems is often defined as the number of variables and/or the quantity or quality of the relations between them. It is a common assumption that these (and other) formal characteristics influence the difficulty in solving the problems expressed in the control performance and the knowledge. In this experimental study the influence of two formal characteristics on problem-solving behavior is investigated. These formal characteristics are the pcs (the proportion of correct solutions) and the number of variables of dynamic systems. In a resulting 2X2-CRF-design four variants of the dynamic system Shop (Strauss, 1993) were constructed. 120 persons were tested. Each of them controlled one of the variants. The dependent variable was the control performance of the person. It was shown that the control performance strongly depended on the pcs, i.e. lower values of the pcs decreased the control performance. On the other hand an increase in the number of variables did not decrease the control performance, if the pcs was kept constant. The conclusion is that the number of variables is not an adequate measure of complexity in dynamic systems.