Abstract
Solving network selection problem (NSP) for next generation networks is very challenging. This is further compounded by the availability of myriad of choices for the user to choose from in terms of which network to connect to. Furthermore, in 4G environment, mobility scenarios are highly dynamic and current approaches tends to ignore this. Additionally, these solutions tend to focus on solving one aspect of the NSP. This paper proposes a holistic and dynamic network selection solution based on gray relational analysis (GRA) and Analytical Hierarchical Process (AHP). In order to correctly assess the context on which the network selection occurs, our method uses attributes collected from the three main stakeholders of a NSP mainly: user, mobile node and the candidate networks. Depending on the type attribute, the proposed method uses a combined method of mean value of past sessions as well as most recent value as means to collect the attribute value. This is so that past as well as current scenario is taken into consideration. Weights are assigned to these attributes indicating their relative importance. Previous approaches mostly uses quality of service requirements (QOS) of the executing application as a guide for indicating the weight. The proposed scheme not only uses QOS requirements but also combines user preference and mobile node's mobility scenario into consideration in determining the weights. Finally, ranking of the candidate networks is done using GRA.