Multi-agent British Sequential Auction Sorting Strategy

Auctions are now being widely studied as a means of buying/selling resources in multi-agent systems. Existing work studied auctions mostly based on combinatorial auctions, and the most obvious problem that bids on combinations of items impose is in selecting the winning set of bids. Call this the Combinatorial Auction Problem(CAP). CAP can be formulated as an NP-Complete Program, so there are not largest application domains for combinatorial auctions. S.S.Fatima et al. built up a multi-objects-auction model on the basis of the model put forward by J.K.Goeree and T.Offerman, considering common value and private value at the same time; then they studied the optimum auction agenda, which is based on the auctions that are yet to be conducted.In many real-world cases bidders consider both auctions yet to be conducted and have been conducted. Given this, we analyze two objects sequential English auctions, show that optimal agenda may be either by increasing order or decreasing order of the dispersion of the order statistics of the surpluses. When two objects are complementary, optimum auction agenda is not according to S.S.Fatima et al.'s theory.In order to determine optimum auction agenda, we put forward OASA arithmetic, which is based on auctioneer's total revenue. This arithmetic can rapidly get optimal agenda from n! schemes.At last, we constructed a simulated sequential English auction system; and made two experiments, one displays that it may be the optimal agenda either by increasing order or decreasing order of the dispersion of the order statistics of the surpluses, the other validates that OASA arithmetic is feasible and effective.