基于Agent的报价学习对碳排放权拍卖的影响

Influence of Agent-based bidding learning on carbon emission rights auction

针对碳排放权序贯拍卖中竞标主体的报价学习行为对拍卖结果的影响,引入自适应性报价模型并改进PSO(particle swarm optimization)智能算法来模拟控排企业报价动态变化的学习机制,并构建基于多Agent的拍卖仿真模型,实验分析不同拍卖情景下,报价学习行为对碳排放权拍卖出清价格、年均履约水平以及拍卖效率的影响.实验结果表明:报价学习行为可以显著提高企业年均履约水平,但是在歧视价格拍卖以及统一价格拍卖(市场供小于求时)下,会显著降低碳排放权出清价格和拍卖效率;歧视价格拍卖的出清价格和拍卖效率都高于统一价格拍卖,但从公平性以及降低企业成本的角度出发,统一价格拍卖可以减少拍卖对没有学习行为的竞标主体的不公平性以及避免报价学习行为对拍卖效率的影响.

This paper introduces a self-adaptive pricing model and improved particle swarm optimization(PSO) algorithm to simulate the learning mechanism of the dynamic change of bidding prices to examine the impacts of bidders’ learning bidding strategies on carbon allowance sequential auction. A multi-agent auction simulation model is constructed to study the impacts of bidders’ learning behaviors on carbon allowance clearing prices, annual performance level, and auction efficiency under different experimental situations. The results show that the bidding learning behaviors can significantly increase the average annual compliance level of bidding firms but reduce the clearing prices and auction efficiencies under discriminatory-price auction and uniform-price auction(when market supply is less than demand). Meanwhile, discriminatory-price auctions have higher clearing prices and auction efficiency than uniform-price auctions. However, in view of fairness and reducing corporate costs, uniform-price auctions reduce unfairness for firms who do not have bidding learning behaviors, and avoid the influence of bidding learning behaviors on auction efficiency.