Abstract

A large portion of expenses in the forest industries is associated with wood supply procurement. Numerous suppliers are involved and securing wood supply contracts with competitive prices is a constant challenge for procurement managers. A major difficulty is the procurement exposure to various sourcing risks including the start of the spring thaw, contract breach, or unreliability of suppliers. A procurement plan should anticipate random events and include measures that counter their negative impact. Recourse actions must be planned by considering volume uncertainty and wood price fluctuations. Relying on manual tools is hardly capable of considering all aspects of this problem. A stochastic programming approach is proposed to support the development of a procurement plan. In this model, several types of contracts including fixed, flexible and option contracts with different durations are included. The proposed selection of contracts from a stochastic programming model yields average optimality in the presence of plausible scenarios. The developed two-stage stochastic programming model decides on the selection of the optimal portfolio of contracts to minimize total procurement costs. Based on a case study in Quebec, an average saving of 4% was shown by using stochastic programming compared to the deterministic approach.

摘要

森林工业的很大一部分费用与木材供应采购有关。涉及众多供应商，以具有竞争力的价格获得木材供应合同是采购经理面临的持续挑战。一个主要困难是采购面临各种采购风险，包括春季解冻、合同违约或供应商不可靠等。采购计划应预测随机事件，并包括应对其负面影响的措施。必须通过考虑数量不确定性和木材价格波动来计划追索行动。依靠手动工具很难考虑到这个问题的方方面面。提出了一种随机规划方法来支持采购计划的制定。在这个模型中，几种类型的合同包括固定的、包括具有不同期限的灵活和期权合约。建议从随机规划模型中选择合同，在存在合理场景的情况下产生平均最优性。开发的两阶段随机规划模型决定最佳合同组合的选择，以最小化总采购成本。根据魁北克的一项案例研究，与确定性方法相比，使用随机编程平均可节省 4%。