Improved Lower Bound for Competitive Graph Exploration
We give an improved lower bound of 10/3 on the competitive ratio for the exploration of an undirected, edge-weighted graph with a single agent that needs to return to the starting location after visiting all vertices. We assume that the agent has full knowledge of all edges incident to visited vertices, and, in particular, vertices have unique identifiers. Our bound improves a lower bound of 2.5 by Dobrev et al. [SIROCCO'12] and also holds for planar graphs, where it complements an upper bound of 16 by Kalyanasundaram and Pruhs[TCS'94]. The question whether a constant competitive ratio can be achieved in general remains open.
READ FULL TEXT