Abstract
City authorities monitor emissions in critical regions of the city. If the emissions reach a certain threshold, city’s traffic management (TM) restricts vehicles’ access to these areas or provides alternative traffic routes. Therefore, TM changes the city’s traffic strategy. These changes impact
the delivery routing for courier, express, and parcel services (CEPs). Each strategy induces a travel time matrix between the CEP’s customers. For efficient routing, CEPs may anticipate potential future strategy (i.e., matrix) changes in their routing decisions, e.g., derived from historical
observations. Nevertheless, the TM is able to predict future strategy changes for a limited time horizon. Hence, it may be beneficial to actively communicate future strategy changes to the CEPs. This may enable CEPs to route their delivery vehicles efficiently and may reduce the overall city traffic and emissions. In this paper, we analyze the impact of anticipation and communication of future strategy changes. We model the described problem as a dynamic vehicle routing problem with stochastic transitions of travel time matrices and apply anticipatory methods of approximate
dynamic programming. In a case study for the city of Braunschweig in Germany, we show the advantages of both anticipatory dynamic routing and communication between TM and CEPs.
the delivery routing for courier, express, and parcel services (CEPs). Each strategy induces a travel time matrix between the CEP’s customers. For efficient routing, CEPs may anticipate potential future strategy (i.e., matrix) changes in their routing decisions, e.g., derived from historical
observations. Nevertheless, the TM is able to predict future strategy changes for a limited time horizon. Hence, it may be beneficial to actively communicate future strategy changes to the CEPs. This may enable CEPs to route their delivery vehicles efficiently and may reduce the overall city traffic and emissions. In this paper, we analyze the impact of anticipation and communication of future strategy changes. We model the described problem as a dynamic vehicle routing problem with stochastic transitions of travel time matrices and apply anticipatory methods of approximate
dynamic programming. In a case study for the city of Braunschweig in Germany, we show the advantages of both anticipatory dynamic routing and communication between TM and CEPs.
