A Probabilistic Approach for Rendezvous Decisions with Uncertain Data

In this paper, the p-Probabilistic Rendezvous algorithm (in short p-PR) is introduced. The main difference between the traditional rendezvous and ours is that the former requires exact location information to calculate the rendezvous. Without this restriction, p-PR is more suitable in applications, where the uncertainty of the locations is commonplace. The main novelty of our approach is that the geometric properties of the rendezvous problem are exploited to avoid the exhaustive examination of all combinations of the object instances. Experiments using synthesized and real-world datasets verified the efficiency and scalability of p-PR. It achieves a time complexity of O(slog(s)), where s is the sum of all the instances of all objects.