|Modelprojekt, 1. modul, 2003, id:285|
|Vejleder:||Bernhelm Booss Bavnbek|
|Findes på RUb:||Ja|
The focal point of this project is routing in mobile ad-hoc networks. The original purpose was, to make possible an analytic comparison of real-life routing algorithms. This turned out to be difficult, and we have therefore worked with a broader question: What can mathematical models express, concerning routing in mobile ad-hoc networks, and what deficiencies does these models have with regard to an assessment of real-life routing algorithms ? We treat models based on stochastic graph theory. In particular we treat in depth two models. One which describes the probability that a finite network is connected, and another which describes how utilization of multiple routes, can decrease routing overhead by delaying the time, at which point new routes must be searched for. The conclusion is, that it is limited what we at present moment can say about routing, with the aid of mathematical models. However, as a start, we can give a few limited partial results. The studied models have several deficiencies, among others, the lack of a time aspect, and lack of named nodes. Furthermore it is questionable whether some of the assumptions, used in the models, are too restrictive.