Adaptive Routing Based on Bayesian Network and Fuzzy Decision Algorithm in Delay-Tolerant Network

Abstract

The topology of delay-tolerant network (DTN) changes frequently. The whole network is disconnected and there exists no end-to-end path for messages. We need a kind of mechanism called store-carry-forward to overcome this fault in DTN. However, traditional end-to-end routing protocols do not meet this requirements. Researchers developed many kinds of DTN routing protocols and different routing protocol fits different application characteristics. Because of the variable application characteristics, we should choose suitable routing protocols according to application requirements. For example, if the current traffic is heavy or the size of the message is too big, we should use routing protocols with small copy number; if the transmission is emergent, we should use routing protocols with short delay. This paper proposes the concept of adaptive routing, meaning choosing routing protocols based on application characteristics in DTN. We develop an adaptive routing scheme according to particular requirements, history network information and current network conditions in network scenarios and validate that with adaptive routing scheme the network performances improve a lot in many aspects, such as delay, overhead and delivery through simulations.

Relay Node Selection based on Adaptive Neuro- Fuzzy Inference System in Delay Tolerant Networks

“…To guide the forwarding process, nodes can evaluate their relationships with each other, in terms of "closeness", which summarizes both temporal and special information, based on contact history. In the [25], is introduced an adaptive routing based on Bayesian network and fuzzy decision algorithm in delay-tolerant networks. By now, it is designed different kinds of routing protocols for DTNs networks .An approach has been developed in [26]which is Neuro-Adaptive Learning Fuzzy-Based System for Actor Selection in Wireless Sensor and Actor Networks.…”

Vehicular Delay Tolerant Network Routing Algorithm Based on Bayesian Network

“…Liang et al [18] proposed the concept of adaptive routing based on BN according to application characteristics in DTN, which means that before transmitting data, a suitable routing protocol is decided to use in order to meet application requirements and to achieve better network performances. However, their BN model focuses on the choice of different routing protocols, not on the scheme of message forwarding.…”

“…NB classification model is a simple and effective classification model, but its conditional independence assumption makes it unable to express the dependency among attributes and reduces the classification accuracy. On the contrary, the Bayesian Network (BN) [18] represents a set of attributes and their conditional dependencies via a probabilistic graphical model, thus can improve the accuracy of inference significantly. According to our previous work [19], a routing algorithm based on BN was proposed for VDTN, which utilizes the BN model to preserve the dependency of the attributes of node (e.g., location, time, contact, etc.)…”