基于高斯分布的在线采购机制设计及竞争分析

依附于互联网电子商务的在线采购拍卖交易, 对传统的贝叶斯离线拍卖理论提出新的挑战, 因为面对不同时间点的投标, 采购电商必须即可决策出是否中标以及购买价格。鉴于此, 对于诸如石油、煤、粮食等无限可分商品的电子采购, 本文基于投标具有高斯分布特征设计了一种激励相容的在线采购策略, 演绎出在线采购的数学模型, 利用Runge-Kutta数值算法, 通过Matlab编程求解出采购电商在线定价策略的需求曲线及其对应的竞争比, 最后, 利用数值模拟, 将在线采购机制策略与纯竞争分析得到的在线采购策略比较, 结果显示利用了高斯分布信息的在线采购策略的竞争性能由于利用了投标的统计信息而得到了提高。

Design of Online Purchasing Mechanism Based on Gaussian Distribution and Its Competitive Analysis

Purchasing auction attached to Internet e-commerce poses new challenges to the traditional Bayesian offline auction theory,because in the face of bidding at different time,the purchasing transactions mechanism of e-commerce must be able to determine whether to win the bid or the purchase price.In view of this,for e-procurement of infinitely separable commodities such as petroleum and grain,this paper designs an incentive-compatible online procurement strategy based on the bids with Gaussian distribution characteristics,and deducts the mathematical model of online procurement.We give an algorithm using Runge-Kutta numerical method to solve the complicated mathematical model.With the numerical simulation,the online procurement mechanism strategy is obtained.The competitive performance of the online procurement strategy using the information of bids with Gaussian distribution is much better than the online reverse auction strategy obtained by pure competition analysis.In summary,the use of statistical information in the design of competitive strategies for online procurement auctions significantly improves the competitive performance of online strategies.