The modelling and analysis of command and control decision processes using extended time petri nets.

Abstract

Effective command and control is crucial to both military and non-military environments. Accurate representations of the processes associated with the inter and intra activities of nodes or agencies of such systems is essential in the analysis of command and control. One of the most important things is to be able to model the decision processes. These are the parts of the system that make decisions and then guide the direction of other elements in the system overall. This thesis uses a new type of extended time Petri net to model and analyse command and control decision processes. A comprehensive review of existing time Petri net structures is given. This concludes with the introduction of a time Petri net structure that incorporates the most commonly used time structures. This extended time Petri net structure is then used in the definition of the basic modelling blocks required to model command and control decision processes. This basic modelling block forms the basis of the direct analysis techniques that are introduced in the thesis. Due to the transient nature of the systems being modelled and the measures of interest a new type of measure is introduced, the mean conditional first hitting reward. This measure does not currently appear to be part of the stochastic process literature. Explicit procedures are given to determine the hitting probabilities and mean conditional first hitting reward for decision process models and discrete, continuous and semi-Markov chains. Finally the some extensions of the decision process sub-class are considered.