Strong Duality for a Multiple Good Monopolist
summary
Proves strong duality for the single-buyer multi-item revenue problem by mapping the primal (maximize a convex 1-Lipschitz utility functional) to an optimal-transport dual that allows mass-splitting preserving the mean. The dual is tight, so every optimal mechanism has a transportation-flow certificate checkable by complementary slackness. Corollaries: grand bundling is optimal iff two stochastic-dominance conditions between explicit signed measures of F hold; examples show bundling can beat separate pricing (uniform [1,2] yields $2.25 vs $2) and randomized menus can strictly beat any deterministic mechanism.
AI-written summary · about