Towards System-Aware Routes

Within this paper, we focus on the development of algorithms capable of providing route suggestions to multiple simultaneously traveling users such that the (negative) impacts of the routes among each other are minimized. We refer to this kind of routes as system-aware routes. E.g., if multiple travelers are concurrently following the same route in a street network, it is most likely that traffic congestion occur. If the same set of travelers is, however, distributed over multiple roads/routes, the traffic flow is maintained and negative effects like increased CO2 emissions are reduced. Today’s routing and navigation devices are, in most cases, not connected with each other, meaning that each user request is handled separately—even if the routing requests are all processed on a centrally operated infrastructure. However, dependent on the availability of traffic data, some devices adapt route suggestions based on observed (or typical) traffic conditions. This behavior leads to a so-called user equilibrium (UE) [5] where travelers cannot reduce their individual travel time by choosing another route. In contrast to an UE, a system optimum (SO) [5] consists of a traffic state where the average travel time is minimized. Obviously, this statement is equivalent to the observation that in a SO the overall travel time, i.e. the sum of all travel times, is minimized or the (overall) traffic flow is maximized. Negative environmental impacts caused by traffic congestion are reduced as well. Reaching a SO requires cooperative travelers as well as supporting technologies sharing—among others—(traffic) information with each other. First solutions to the basic underlying problem of finding a shortest route in a given network reach back to the late 50ies [2]. There were significant improvements in the last decade, especially with respect to speeding-up techniques via pre-processing, cf. [3] or [1] for a more comprehensive survey,. There is, however,