Abstract
The useful navigation guidance is favorable to considerably reducing navigation time. The navigation problems involved with multiple destinations are formulated as the Directed Steiner Tree (DST) problems over directed graphs. In this paper, we propose a deep learning (to be exact, graph neural networks) based approach to tackle the DST problem in a supervised manner. Experiments are conducted to evaluate the proposed approach, and the results suggest that our approach can effectively solve the DST problems. In particular, the accuracy of the network model can reach 95.04% or even higher.
Recommended Citation
Yen, Benjamin and Luo, Yu, "Navigational Guidance – A Deep Learning Approach" (2021). ICEB 2021 Proceedings (Nanjing, China). 34.
https://aisel.aisnet.org/iceb2021/34