Information Diffusion and Revenue Optimization in Distribution Market

Intermediaries refer to agents that facilitate the resource allocation under certain constrain conditions. In a distribution market, buyers and sellers often interact via intermediaries and pay commissions to them. Due to the inadequacy of information spreading, a commodity can only be allocated to some agent who is close to the seller, which not only decreases the economic efficiency but also limits the seller’s revenue. To tackle these two problems simultaneously, which is commonly a difficult objective in multi-agent systems, we use mechanism design theory and graph algorithms to develop a new auction protocol called Multi-Level Diffusion Mechanism (MLDM). Even when there are costs for assisting in others’ trades, this mechanism can encourage every intermediary to help spread the selling information, either or not she has personal preference of the commodity. We show MLDM performs better than the second price auction both in theory and in numerical simulation no matter how agents’ valuations and commissions are distributed.