We consider a new search paradigm, in which two nodes simultaneously conduct an expanding ring search for a path between each other. The new method generalizes expanding ring search, which is a well-studied flooding-based technique for discovering paths and resources in ad hoc networks. The idea behind the two-sided search is that all nodes that receive a query cache the shortest path to the source of the query; it is therefore only necessary for one node to find a cached route to the other. A two-sided search offers ways to reduce expected costs in ways that a search by one node alone cannot, and as such it weakly dominates a one-sided search. We show how to derive the sequence of search ranges with the minimum expected total search cost in polynomial time. We further consider optimal strategies under time constraints and study various properties and benefits of the two-sided search.
展开▼
