Correlation-robust analysis of single item auction

摘要: We investigate the problem of revenue maximization in single-item auction within new correlation-robust framework proposed by Carroll [2017] and further developed Gravin Lu [2018]. In this auctioneer is assumed to have only partial information about marginal distributions, but does not know dependency structure joint distribution. The auctioneer's evaluated worst-case over uncertainty possible distribution.For optimal design correlation robust-framework we observe that most cases admit a simple form like celebrated Myerson's for independent valuations. analyze compare performances several DSIC mechanisms used practice. Our main set results concern sequential posted-price mechanism (SPM). show SPM achieves constant (4.78) approximation mechanism. also symmetric (anonymous) case when all bidders same distribution, (i) has almost matching worst-correlation as any second price with common reserve price, (ii) number large, converges optimum. addition, extend some on computational tractability lookahead auctions framework.