Online Bipartite Matching with Predicted Degrees

We propose a model for online graph problems where algorithms are given access to an oracle that predicts the degrees of nodes in the graph (e.g., based on past data). Within this model, we study the classic problem of online bipartite matching. An extensive empirical evaluation shows that a greedy algorithm called MinPredictedDegree compares favorably to state-of-the-art online algorithms for this problem. We also initiate the theoretical study of MinPredictedDegree on a natural random graph model with power law degree distribution and show that it produces matchings almost as large as the maximum matching on such graphs.