In system-aware optimization the goal is to apply optimization techniques on system-wide challenges, i.e., challenges which are relevant to multiple stakeholders. In the field of transport optimization one of these challenges is to smooth the urban transportation system. This includes, among others, the minimization of congestion. In this context, (partly contradicting) objectives like the reduction of traffic noise or the maximization of livability of a city have to be considered.
The overall research approach at AIT’s Center for Mobility Systems is to come up with a holistic view on the transportation system. In the field of system-aware optimization this means that multiple current optimization problems are tackled at the same time. E.g., when talking about route planning the standard approach is to calculate a route for one traveler which is optimized with respect to some objective function such as travel time, energy consumption or any other criterion. However, the impacts of the therewith proposed route on other transport users are not considered at all. This is relevant in particular when multiple transport users are behaving similarly. E.g., typical congestion during peak hours is due to an individual optimization approach, i.e., egoistic behavior.
In system-aware optimization, the approach is now to tackle all individual optimization problems (e.g. route planning) concurrently. Especially, the impacts of individual decisions on the other optimization problems are considered and integrated into the optimization approach. Therefore, the multiple concurrent optimization problems are merged into one multi-stakeholder, multi-objective optimization problem where the overall goal is to find a system-wide optimum.
AIT applies system-aware optimization techniques especially in the field of traffic flow optimization by an integrated multi-stakeholder routing approach where individual routes are computed concurrently. The impacts of individual routes on the routes of other transport users are estimated via adapted fundamental diagrams of traffic flows such that the extra travel time (due to congestion) can be sized. Application areas are traffic flow optimization in transport networks as well as optimization of routes for highly-automated vehicles in closed or semi-closed surroundings like train stations or production facilities.
