(The) extension of activity cycle diagram and its generality

Abstract

The simplicity of the two-symbol activity cycle diagram (ACD) has made it easy to learn and use to describe discrete event systems in a natural way. But as the complexity of the model increases, the ability of the ACD to provide a full description of the system under consideration becomes more limited. To tackle down the obstacle of ACD, this thesis proposes two classes of extensions to the classical ACD: 1) the extended ACD and 2) the parameterized ACD. The extended ACD is an extension with arc attributes (condition and multiplicity) so as to increase the modeling power of the ACD. The parameterized ACD is another extension with parameterization on the activity and queue of the ACD and token attribute so as to increase the ease of model building and to reduce the model complexity. The generality of proposed extended ACD is proved with two approaches, which has not yet been done on the ACD: 1) conversion of the extended ACD into a Turing-equivalent modeling formalism, extended Petri Net; and 2) creating the extended ACD model of a Turing machine. This implies that no other extension except the arc condition is necessary to increase the modeling power of the ACD. Other extension may increase the ease of modeling or reduce the model complexity. The comparative study is also provided to show that the ACD is a good modeling formalism in qualitative and quantitative ways. In a qualitative analysis, it is compared on the six criteria of selecting a modeling formalism with the Event graph, which also allows the parameterization on event vertices and has a modeling power equivalent to a Turing machine. In a quantitative analysis, the simulation model complexities between ACD and event graph models are measured. This comparative study shows that the ACD is a good modeling formalism in that the state changes are cleared defined through the relationship between activities and queues. Thus, this reduces the complexity of each node; in the meanwhile, the structural comple...