Cyber Physical Regional Freight Transportation

The purpose of this research is to develop optimization and control techniques and integrate them with real-time simulation models to achieve load balancing in complex networks.  Our application case is the regional freight system.  Freight moves on rail and road networks which are also shared by passengers.  These networks today work independently, even though they are highly interdependent, and the result is inefficiencies in the form of congestion, pollution, and excess fuel consumption.  These inefficiencies are observed for example by the peaks of demand across time and space.  Inefficiencies exist in part due to lack of information and appropriate tools, and in part due to lack of policies and institutional structures that would promote more integrated operations. 
The fundamental concept of this research is freight load balancing:  what is the best (most efficient and sustainable) allocation of freight shipments across rail and roadway? The proposed freight load balancing system is a classical Cyber physical system where communications, controls, computations and optimization strategies are integrated to move freight transportation practices to higher levels of efficiency and sustainability. We propose a co-simulation optimization control approach for controlling and optimizing systems of large dimension and complexity by replacing simple models with more accurate simulation models in a feedback control/optimization loop. The method will be used to design and evaluate a scalable CPS dynamic freight load balancing system whose objective is to optimize freight loads across two networks, rail and road also shared by passengers, by taking into account their temporal and spatial characteristics. We will use real time data together with a simulation testbed to validate and evaluate our approach, and collaborate with key stakeholders in the development of our models in order to promote industry testing and implementation.  We will address policy and institutional constraints as well as current practices and level of expected coordination in using the proposed load balancing system. The success of the co-simulation optimization approach for freight load balancing is a first step in improving performance of many other complex networks.