考虑不确定履约率的全局鲁棒供应组合决策研究

生产和采购的全球化为供应链企业带来了更多的复杂性与不确定性。收集全面信息和使用有效的工具是控制决策风险、减少不确定性危害的有效途径。在供应商选择与订单分配管理中,供应商履约率是影响企业供应组合决策的主要因素,如何对其进行定量刻画是迫切需要解决的问题。已有的研究通常是由期望方法或鲁棒方法来刻画履约率。然而,期望履约率法往往忽略了履约率的波动性,从而造成需求损失;而鲁棒履约率法通常存在过于保守、退化过于迅速两个缺点,使企业承担不必要的成本。本文研究供应商选择与订单分配决策问题,使用全局鲁棒优化的两个不确定集合来刻画供应商履约率,同时给出了这一问题的确定性鲁棒表达和一种多项式时间求解算法。最后,通过仿真实验证实本文的全局鲁棒优化模型可克服期望模型和一般鲁棒模型的缺点。本文提出的算法能有效求解本文全局鲁棒优化等价确定性问题,为企业的供应商选择和订单分配决策提供更为灵活精确的辅助工具。

Globalized robust supply portfolio problem with uncertain reliability

The globalization of production and procurement process has become one of the main trends of today′s global economy.Yet,the globalization is a complicated and controversial issue.While bringing lower prices and more choice of raw materials for procurement managers,the globalization also brings complexities and uncertainties to the decisions of supply chain managers,such as supply reliability uncertainty,cargo damage uncertainty,forward price uncertainty and so on.Gathering more general information,and using more effective decision support tools are two significant ways to control uncertainty risks.Therefore,supply chain managers should collect more data and use effective decision support tools in their work.Since procurement management is one of the most important issues of enterprise supply chain management and supply reliability is a key factor influencing the enterprise supply portfolio decision,we study the supplier selection and order allocation problem with supply reliability uncertainty in this paper.Additionally,we propose a practical decision model for entertainment’s procurement managers in their supply portfolio decisions considering supply reliability uncertainty.And a polynomial time algorithm for the mathematical model is also provided.The model and algorithm proposed in this paper could serve as a decision support tool for entertainment’s managers in their supply portfolio decisions for procurement.When dealing with the problem of supplier selection and order allocation,the procurement manager who faces many available supply contracts provided by the supplier aims to reach the supply portfolio decision that fully meets the demand at the lowest cost.In this process,the procurement manager faces both demand and supply uncertainties.In existing studies,the uncertainty demand is always characterized by expectation.And two basic approaches have been used to characterize the uncertain supply reliability:expectation formulation and robust formulation.Despite its widely use,the expectation formulation often neglects the volatility of supply reliability,which makes the firm suffer from costs of unmet demand and excess inventory.Robust formulation,although overcomes the shortcomings of expectation formulation,often results in unnecessary cost of emergency shipment due to its two shortcomings:over conservation and over-quick degeneration.In order to overcome the disadvantages of the above models,we propose a globalized robust model for supply portfolio problem in this paper.In the globalized robust model,supply reliability is formulated by two uncertain sets.The globalized robust model requires full feasibility for supply reliability values in the inner uncertainty set analogous to the robust model.However,infeasibilities are allowed with supply reliability values in the outer uncertainty set,where the violation is controlled by the distance of the parameter value from the inner uncertainty set.Therefore,the globalized robust model better reflects volatility than the expectation formulation does.Moreover,the globalized robust model overcomes the two drawbacks of robust formulation.In another words,the mathematical model proposed in this paper is more flexible and advantageous than the those in existing research,which is much more helpful for the practical application in supply portfolio decision.However,the globalized robust model is a difficult semi-infinite programming,which can’t be solved directly by existing mathematical programming methods.To cope with this difficulty,by using Lagrangian duality theory,we propose an equivalent deterministic counterpart of the globalized robust model.Moreover,we provide the technical proof of the equivalence between the deterministic of globalized robust supply portfolio model and the complex semi-infinite globalized robust model.In this way,we transform the complex semi-infinite globalized robust model into a deterministic nonlinear convex optimization with nonlinear coneconstrains.However,since the deterministic counterpart of globalized robust supply portfolio model is nonlinear convex optimization with nonlinear cone-constrains,it is still hard to solve the model by existing methods.To this end,by using regularization and linearization augmented Lagrangian technique,we provide a polynomial time algorithm for the deterministic counterpart of globalized robust supply portfolio model.In this way,the advantageous globalized robust model proposed in this paper is polynomial time solvable.Therefore,the model and algorithm proposed in this paper are possible to serve as a decision support tool for entertainment’s procurement managers in their supply portfolio decisions.Finally,to confirm the advantages of globalized robust portfolio model and the efficiency of the algorithm proposed for the deterministic counterpart model,we conduct simulation experiments.Our simulation experiments show that the proposed model does overcome the shortcomings of the expected model and the robust model.Moreover,the algorithm proposed in this paper can solve the counterpart model of globalized robust supply portfolio problem effectively.The model and algorithm proposed in this paper can serve as a useful tool for entertainment’s managers in their supply portfolio decisions.They are helpful for the practical application in supply portfolio decision.