Adaptive Configuration of Pervasive Computing System with QoS Consideration

User mobility in pervasive computing makes it fundamentally different from traditional desktop computing. Accordingly, the system configurations, such as resource allocation and routing, are extremely dynamic since available resources and users´ needs may change when users are moving around different locations. There have been extensive works on the resource allocation for the users´ needs. However, previous works did not consider complex situations, such as conflicting relations between different applications. This paper models a pervasive computing system from three parts: available network resources, available applications, and conflicting relations of the applications. To configure the system according to the dynamical changes, our work focuses on how to find the best set of applications, under the network resource and QoS routing constraints, to achieve the maximum system benefit for the users. We first propose an optimal solution using integer linear programming (ILP) scheme. Then a simple and effective heuristic is presented for better scalability. Extensive simulation results demonstrate that our ILP and heuristic achieve very good performance.