Continuous-time production, distribution and financial planning with periodic liquidity balancing
Continuous-time production, distribution and financial planning with periodic liquidity balancing
复制标题
DOI：
10.1007/s10951-016-0488-7
复制
发表时间：
2017-06
影响因子：
2
通讯作者：
W. Albrecht;M. Steinrücke
中科院分区：
文献类型：
--
作者：
W. Albrecht;M. Steinrücke
由于对核心竞争力的不可避免的关注，即使是中小型公司也越来越多地被迫形成供应链（SC）网络。然而，他们的具体情况往往是缺乏股本，进入资本市场的机会有限，因此必须利用银行贷款来启动生产和销售。在短期的多天计划范围内，运营和财务都必须精确地安排，以便为每个网络合作伙伴和网络管理人员提供实用的指导。为此，需要进行连续时间建模。此外，为防止破产，有必要协调特定地点的业务活动和全网络的金融交易所产生的货币后果。由于银行透支可用于克服短期内的财务失衡（例如，几天甚至几小时），应确定适当的流动性管理时间间隔。这些间隔的实现需要离散时间建模。在这种情况下，主要的挑战是将上述两种建模技术联合收割机组合在一个公共决策模型中。为了解决这个问题，一种新的混合整数非线性规划（MINLP）的开发，这使得精确的规划和调度的供应链操作以及相关的金融交易，一方面，和定期的流动性平衡的另一方面。数值分析是基于随机生成数据的测试场景。正如我们发现的那样，即使是MINLP的小问题实例，例如，一个三阶段的供应链，在每个阶段有三个网站，不计算高性能的硬件和商业非线性标准求解器，我们另外提出了一个等效的线性化版本的决策模型。后者可以在可接受的计算时间内使用CPLEX求解器进行优化。
Due to the inevitable focus on core competencies, even small- and medium-sized companies are increasingly forced to form supply chain (SC) networks. However, their specific situation is often characterized by a lack of equity and limited access to capital markets, so that bank loans must then be used to initiate production and distribution. Within a short-term multi-day planning horizon, both operations and finance must be scheduled precisely in order to obtain practical instructions for each network partner and the network managers. For this purpose, continuous-time modeling is required. Additionally, a coordination of monetary consequences resulting from both site-specific operational events and network-wide financial transactions is necessary to prevent insolvency. As bank overdrafts can be used to overcome financial imbalances during short periods (e.g., days or even hours), appropriate time intervals for liquidity management should be determined. The implementation of these intervals requires discrete-time modeling. In this context, the main challenge is to combine both of the aforementioned modeling techniques within a common decision model. To address this problem, a novel mixed-integer nonlinear program (MINLP) is developed, which enables exact planning and scheduling of SC operations as well as related financial transactions on the one hand, and periodic liquidity balancing on the other hand. A numerical analysis was based on a test scenario with randomly generated data. As we found out that even small problem instances of the MINLP, e.g., a three-stage supply chain with three sites in each stage, were not computable with high-performance hardware and a commercial nonlinear standard solver, we additionally propose an equivalent linearized version of the decision model. The latter could be optimized within acceptable computation time using the CPLEX solver.