1 paper
Michael Kapralov, Ian Post, Jan Vondrak
We prove that no online algorithm (even randomized, against an oblivious adversary) is better than 1/2-competitive for welfare maximization with coverage valuations, unless $NP = R…