管理工程学报2011，Vol.25Issue(2)：220-227,8.
基于Bayes需求预测更新的供应链合作策略研究
A Policy for Supply Chain Coordination with Bayes Demand Forecasting
摘要
Abstract
Demand variety and long replenishment lead time are two common characteristics of seasonal products. The main task of managing supply chains for seasonal products is to minimize overstocks and back orders. Institutionalizing quick response (QR)policies to reduce demand uncertainty is a practical approach. These two major issues are related to QR policy. The first issue is to determine when to order and how much to order from a buyer's perspective. The second issue is to learn how to motivate supply chain partners in order to achieve a win-win situation from the perspective of the entire supply chain.The current literature neglects the effect of lead time on forecasting error and supply chain collaberation. This paper investigates supply chain coordination policies using Bayes demand forecast model for a perishable-product supply chain. We proposed a demand forecasting model to help optimize ordering quantity and lead time. The need for supply chain coordination was discussed based on the proposed model. In addition, we proposed a three-parameter risk-sharing contract to coordinate the decentralization of decisions across the supply chain. Channel coordination conditions were derived from the proposed contract. An optimal search algorithm and a sensitivity analysis of parameters were also provided to illustrate conclusions and procedures. Our test results showed that such a contract can not only maximize supply-chain profits, but also flexibly distributes profits among supply-chain members in proportion to each supply-chain member's contribution.Section 1 discussed three major assumptions related to the management of seasonal products. Seasonal products have high demand uncertainty which can cause inventory and backorders. Retailers often dispose inventory after the season at a discounted price. This paper assumes that retailers only have one ordering opportunity through the entire season. The objective of a retailer's decision is to maximize its profit.Lead time is composed of n mutually-independent components. Additionally, lead time can be crashed with additional operating costs. Information about sales of related items can be used to reduce forecast error for the expected demand. An updated demand forecast can be obtained by utilizing Bayes' theorem. Section 2 formulated solutions to maximizing a retailer's profit by making independent decisions, and the entire supply chain's profit by making integrated decisions. We further discussed the attributes of the proposed optimal models as follows.Lemma 1 The maximum value of the retailer's expected profit Π (q,t) exists. Retailers can benefit from information upates.The maximum value of the entire supply chain's expected profit Ⅱ( q ,t) exists. The entire supply chain can also benefit from information upatos.Lemma 2 Without coordination, the optimal decision to maximize the retailer's profit may decrease profits of the entire supply chain and the manufacturer; that is, ΠDr ≥ ΠCr ,ΠDm ≤ ΠCm,ΠDl ≤Πcl.In contrast, ΠDr,ΠDm and ΠDl are profits of the retailer,the manufacturer and the entire supply chain under decentralized decision scenario; Πrc,Πcm and Πcl are the profits of the retailer, the manufacturer and the entire supply chain under centralized decision scenario.Based on the findings, we further developed an efficient iteration algorithm to locate the optimal solution.Section 3 developed a three-parameter contract to help coordinate supply chain operations by combining buy-back and risk-share contracts. The three-parameter contract can arbitrarily allocate profits of supply chain channels between manufacturers and retailers by tuning a range of w values. Section 4 presented a numerical analysis of the proposed models and algorithms. Our analysis results showed that the algorithm is efficient and the three-parameter contract is effective and flexible to coordinate supply chain operations. In summary, decisions made on the order quantity and lead time have important effect on supply chain parties. Effective collaboration between supply chain partners can obtain the win-win situation.关键词
需求预测/Bayes/信息更新/供应链合作/契约Key words
demand forecast/ Bayes/ information updating/ supply chain coordination/ contract分类
管理科学引用本文复制引用
宋华明,杨慧,罗建强..基于Bayes需求预测更新的供应链合作策略研究[J].管理工程学报,2011,25(2):220-227,8.基金项目
国家自然科学基金资助项目(70872047) （70872047）
