Multiobjective optimization of service delivery over a heterogeneous wireless access system

Abstract

This paper investigates the problem of optimal assignment of wireless interfaces for service delivery in the context of a heterogeneous wireless access system. We formulate service requests initiated by the applications running on the mobile device as a multiobjective optimization (MOO) problem: optimal mapping of application requirements to wireless interface while minimizing the monetary cost and maximizing the satisfaction with respect to QoS metrics. This problem is then solved for three different access scenarios using an exhaustive search algorithm and the NSGA-II MOO algorithm. Our aim is to demonstrate how MOO works in such a context, nevertheless for real life problems where exhaustive search cannot handle the problem, we expect NSGA-II to be a more appropriate tool. The results obtained by these algorithms are then discussed and their quality is compared to a more traditional cost function based on weighted sum.