Methodologies for calculating decision-related event occurrence times, with applications to effective routing in diverse MANET environments

Abstract

To capture the dynamics of context-aware decision making for network functions in Mobile Ad Hoc Networks (MANETs), the paper introduces the concept of Decision-Related Event Occurrence Times (DREOTs). By considering a generic scenario, the paper develops a broadly applicable analysis methodology, which produces upper and lower bounds for the tail probabilities and moments of a DREOT. The nature and effect of the decision is abstracted away to quantities called ‘decision policy impact descriptors’. These are calculated on a case-by-case basis, by considering properties of the specific decision policy and mobility pattern in hand. The power of the approach is demonstrated through a detailed application targeting the analysis of a preexisting routing protocol that bases its next hop forwarding decisions on the location and motion of the neighboring nodes. In the considered application the DREOT represents the time a node retains a message before forwarding it to the next hop and the network nodes move according to the Random Direction mobility model. The proposed methodology is seen to bring more rigorous understanding of the relationship between the retaining time and the relevant context and to yield refined expressions, whose use improves the protocol’s performance for all network conditions in a very diverse set. Furthermore, the broad applicability of the methodology is illustrated through a brief discussion of another application, not related to routing.