Efficient communication in agent-based autonomous logistic processes

Transport Logistic Network; Clustering; Routing; Negotiation. - Transportation of goods plays a vital role for the success of a logistics network. The ability to transport goods quickly and cost effectively is one of the major requirements of the customers. Dynamics involved in the logistics process like change or cancellation of orders or uncertain information about the orders add to the complexity of the logistic network and can even reduce the efficiency of the entire logistics process. This brings about a need of integrating technology and making the system more autonomous to handle these dynamics and to reduce the complexity. Therefore, the distributed logistics routing protocol (DLRP) was developed at the University of Bremen. In this thesis, DLRP is extended with the concept of clustering of transport goods, two novel routing decision schemes and a negotiation process between the cluster of goods and the vehicle. DLRP provides the individual logistic entities the ability to perform routing tasks autonomously e.g., discovering the best route to the destination at the given time. Even though DLRP seems to solve the routing problem in real-time, the amount of message flooding involved in the route discovery process is enormous. This motivated the author to introduce a cluster-based routing approach using software agents. The DLRP along with the clustering algorithm is termed as the cluster-based DLRP. In the latter, the goods are first clustered into groups based on criteria such as the common destination. The routing is now handled by the cluster head rather than the individual transport goods which results in a reduced communication volume in the route discovery. The latter is proven by evaluating the performance of the cluster-based DLRP approach compared to the legacy DLRP. After the routing process is completed by the cluster heads, the next step is to improve the transport performance in the logistics network by identifying the best means to transport the clustered goods. For example, to have better utilization of the transport capacity, clusters can be transported together on a stretch of overlapping route. In order to make optimal transport decisions, the vehicle calculates the correlation metric of the routes selected by the various clusters. The correlation metric aids in identifying the clusters which can be transported together and thereby can result in better utilization of the transport resources. In turn, the transportation cost that has to be paid to the vehicle can be shared between the different clusters. The transportation cost for a stretch of route is calculated by the vehicle and offered to the cluster. The latter can decide based upon the transportation cost or the selected route whether to accept the transport offer from the vehicle or not. In this regard, different strategies are developed and investigated. Thereby a performance evaluation of the capacity utilization of the vehicle and the transportation cost incurred by the cluster is presented. Finally, the thesis introduces the concept of negotiation in the cluster based routing methods. The negotiation process enhances the transport decisions by giving the clusters and the vehicles the flexibility to negotiate the transportation cost. Thus, the focus of ...