Machine-Learning based Transportation Network Sparsification for IoT Trucking Automation and Optimization
Di Chang, Zhiming Wang, Xia Zhang
Abstract
As the rapid manufacturing and industrial development, Internet of things (IoT) has become a key role in AI-smart transportation network especially for smart trucking automation and routing optimization. A maximum spanning k-tree algorithm is proposed to optimize the minimal loss of information from original network G to the spanning k-tree. The proposed network sparsification algorithm re-modelling IoT network in one of major U.S. trucking companies achieves great improvement of over performance. The novel Markov k-tree based network sparsification automates route generation, which practically results in increasing truck velocity, accelerating freight to recover service, creating a flexible inventory of loads-in-transit, re-balancing capacity and freight demand. In summary, it also gives a feasible approach to more practically and safely utilize auto-pilot truck based on a real-time sparsified IoT network.