Joint optimization of transmission schedule and routing for network capacity

Abstract

The capacity of a wireless network depends on the interference encountered by flows in the network as they traverse their routes from source to destination. Link-transmissions along a route must be scheduled so as to avoid interference with other links, while the interference-pattern that scheduling has to accommodate itself depends on the routes selected for traffic flow. Therefore, for efficient use of the network, routing and transmission schedule should be considered together. In this paper, we investigate the joint optimization of routing and transmission schedules for capacity maximization. Under the restriction of a single traffic route for each node-pair (as is usual for ease of operation in wireless networks), the formal problem of determining the best route for each node-pair takes the form of an intractable Integer Program for realistic-sized networks, and is not an option to consider. We show that one can, in fact, begin by allowing multi-path routing, since, in this case, the joint choice of routes and transmission schedule is amenable to a Linear Program formulation of a multi-commodity flow problem. Our method then extracts a single-route solution from the multi-path LP solution by solving a much simpler integer program in which we exploit the LPpsilas dasiapath-filteringpsila properties to obtain tight bounds on the number of paths that appear in the integer formulation. We apply this procedure to an example of an ad hoc network and show that the joint optimization of routing and transmission schedule can be carried out in an efficient manner and that it leads to a higher network capacity than realized by choice of transmission schedule alone with pre-specified routing. Our contributions in this paper are two-fold: (a) we propose a method of maximizing mobile ad hoc network (MANET) capacity by the joint optimization of routing and scheduling and (b) the proposed method can also be used to design MANETs for efficiently supporting a given set of missions.