Abstract: In the decentralized environment with incomplete information,a multi-round auction mechanism was designed to solve the order acceptance and scheduling problem by taking global profit optimization of decentralized system as the objective,in which the manufacturer's productivity was limited and each customer's order was characterized by a deadline.An adaptive asking pricing strategy based on supply-demand relationship was designed to adjust the direction of auction dynamically;the bid problem and the winner determination problem were designed with dominant decisions and clear economic benefits objectives.Simulation results showed that the auction mechanism could significantly increase the manufacturer's profit and the total profit of decentralized systems,and the auction mechanism provided over 92% of system value on average found by centralized planning.

Key words: order acceptance, scheduling, decentralized decision-making, auction, adaptive asking pricing

摘要： 在不完全信息的分散决策场景下,针对加工能力有限但需要承诺交货期的单件订单接受与加工调度问题,以优化分散系统全局收益为目标,设计了一个多轮升价拍卖机制。该机制采用基于供求关系的自适应定价策略,动态调整拍卖方向,并以明确的经济收益为目标进行竞胜标、投标占优决策。数值实验结果显示,与先来先服务的加工策略相比,该拍卖机制能显著提高制造商收益和分散系统总收益,分散系统总收益平均达到完全信息下系统总收益的92%以上。

关键词: 订单接受, 调度, 分散决策, 拍卖, 自适应定价