Abstract
AbstractGiven a bipartite graph G = (U,V,E) with U = {1,…,n}, and a positive budget B v for each v in V, a B-matching M in G is a second-price B-matching if, for every edge uv in M, there is a uw in E so that less than B w edges u′w with u′ < u belong to M. The Second-Price Ad Auction with Binary Bids (B2PAA) consists of, given G and B as above, finding a second-price B-matching in G as large as possible. The particular case of this problem where B v = 1 for all v, denoted as Second-Price Matching (2PM), is known to be APX-hard and there is a 2-approximation for it. We present a way to use this approximation and similar ones to approximate B2PAA. Also, we formalize the idea of a competitive market, present an improved approximation for 2PM on competitive markets, extend the inapproximability results for competitive markets and analyze the performance of an algorithm of Azar, Birnbaum, Karlin, and Nguyen for the online 2PM on competitive markets. Our improved approximation can also be used for B2PAA. Finally, we comment on results derived from computational experiments on variants of our algorithm.Keywordsapproximation algorithmscombinatorial auctionsmatchings
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have