Abstract
Communication in Delay/Disruption Tolerant Networks (DTNs) is a challenge because it presumes the absence of a connected end-to-end path (journey) at the time of sending a message to a destination. An efficient selection of a contact node to forward a message is a key in the routing process. Prediction techniques can be used to assist in routing decisions. In this paper we present a journey predictor for DTN based on a context of public transportation system. The journey predictor is centred on an algorithm that builds a graph of predicted journeys and then selects the best journey to a specific destination. This graph is built from a next contact predictor based on Artificial Neural Networks. Experiments were carried out with real contacts of a quasi-opportunist scenario. The proposal outperformed the MaxProp strategy in the most cases considering the number of messages delivered and the execution time.