On learning repeated combinatorial auctions

In this work, we discuss how an intelligent agent learns in combinatorial auctions. It is well-known that finding the optimal allocation to maximize revenue is NP-complete, because this is a typical form of Set Package Problem (SPP). We introduce a framework of reinforcement learning to combinatorial auctions, and discuss how to obtain intelligence about bidding behavior. We show empirical convergence of knowledge within Q-learning framework. By this result, we target fully automated negotiation systems.