Presentation Information
[B-16-10]A Study on a Graph Traveling Network Driven by Local Information
◎△Keito Yokoyama1, Kazuma Aoyama1, Syota Inoue1, Hiroyuki Ohsaki1 (1. Department of Informatics, Graduate School of Science and Technology, Kwansei Gakuin University)
Keywords:
Traveling Network,Local information-driven movement,Growth-shrinkage dynamics
A traveling network (TN) moves through space using only local interactions, without any central controller, but it has so far been formulated on a continuous plane with coordinate-based geometry. This paper proposes a Graph Traveling Network (GTN) that extends the TN onto a graph using only the local information each leaf obtains from its adjacent vertices, assuming neither coordinates nor an embedding. A GTN chooses where to grow from two local quantities—forward room and contact with the existing body—and moves as a body that has reached its size limit contracts at its tips. Simulations on a grid graph show, using only coordinate-free local rules, that a GTN carries its centroid while keeping its size, that contraction is essential to this motion, and that local growth selection improves its directionality.