Scalability investigations on communication traffic in distributed routing of autonomous logistic objects

Abstract

Abstract-In current transport logistics, routing is usually done centrally. A dedicated routing instance solves the optimisation problem of finding the best solution to handle the current set of orders with the set of available vehicles under constraints such as vehicle utilisation, punctuality etc. Because of the increasing complexity of logistic processes, approaches have been suggested recently which change this centralised routing paradigm towards a distributed approach with autonomous logistic entities (vehicles and goods) deciding on their own. To be able to obtain enough knowledge for reasonable decisions, the logistic entities have to communicate with each other. For this interaction, the information exchange concept DLRP (Distributed Logistic Routing Protocol) has been proposed before. The work presented in this paper will focus on the aspect of scalability of communication in a DLRP scenario. Message flooding is identified as potential challenge for the scalability of DLRP, and intelligent flooding restrictions to the communication traffic are applied.

Autonomous Control of Logistic Processes: A Retrospective

“…To improve the scalability of information processing in large networks, various mechanisms to limit the propagation of transport requests can be added to the DLRP t . These include limitation of the search depth, intermediary assessment of partial routes for segments of the total route, re-using data from previous request propagations (Wenning et al 2009;Wenning 2010), or mechanisms for clustering and regional partitioning of transportation networks (Singh et al 2008).…”

“…For the DLRP t , it could be shown that measures to improve the scalability of information processing can significantly reduce communication volumes in large networks without impairing the logistical achievement of objectives. For that reason, the DLRP t can be scaled to larger networks within the given limits of the communication infrastructure (Wenning et al 2009).…”

Implications of Communication Constraints for the DLRP in Transport Logistics

“…[5], [8]) has shown that the route discovery in DLRP can create a large amount of communication traffic if no measures are taken to limit this traffic. However, the main focus there has been on the overall network load, where the route requests are the main contribution.…”

“…Previous work on the traffic volume in DLRP, as described in [5] and [8], utilized a scenario topology with 18 vertices based on a German road network. This scenario has now been extended to a topology with 40 vertices and a corresponding set of connections.…”

“…Each of these vehicles has the same capacity and travel at the speed of 100 km per time unit. In [5], two methods of restricting the amount of goods-initiated route requests were presented and investigated for the 18-vertex topology: forwarding restrictions based on intermediate route evaluation and the application of fixed, scenariowide hop limits for the route request propagation. These fixed hop limits are now applied to the 40-vertex scenario.…”