An agent-based fuzzy constraint-directed negotiation model for solving supply chain planning and scheduling problems
An agent-based fuzzy constraint-directed negotiation model for solving supply chain planning and scheduling problems
复制标题
DOI：
10.1016/j.asoc.2016.07.030
复制
发表时间：
2016-11
期刊：
影响因子：
--
通讯作者：
Chia-Yu Hsu;Bo-Ruei Kao;Van Lam Ho;Lin Li;K. R. Lai
中科院分区：
文献类型：
--
作者：
Chia-Yu Hsu;Bo-Ruei Kao;Van Lam Ho;Lin Li;K. R. Lai
供应链计划与调度问题要求制造商（项目经理）在考虑制造商和供应商之间各种生产约束的情况下，确定产品配置，选择供应商（承包商）进行任务分配，并对项目任务进行调度。大多数相关研究都集中在基于供应链中每个企业提供的完整信息找到最佳解决方案。然而，由于供应链过程的完全分布式性质，完全信息共享的实际实施是困难的，如果不是不可能的话。本文提出了一种基于Agent的模糊约束导向协商（AFCN）模型来解决与供应链计划和调度相关的问题，这些问题被建模为一组通过Agent间约束相互关联的模糊约束满足问题（FCSP）。为了适应供应链中每个企业的观点和利益，FCSP之间的冲突使用AFCN协议通过有限的信息共享和不侵犯隐私的情况下反复交换报价/还价来解决。提议的提议/反提议不仅代表一套可接受的解决办法和对一项行动任务的偏好，而且还代表在这一领域发生冲突的可能性。对于每个FCSP，要约/还价评估的增量过程消除了冗余和不可行的解决方案。代理之间有限的非战略敏感信息的共享，使他们能够阐明他们的对手的意图，通过迭代谈判，使代理可以达成协议，同时确保解决方案的项目规划和调度问题是令人满意的。AFCN模型还具有足够的灵活性，可以在各种生产环境中采用不同的谈判策略，如竞争策略、双赢策略和合作策略。在此，进行了数值研究，以检查所提出的AFCN模型的实际可行性和有效性。实验结果表明，所提出的AFCN模型不仅可以产生一个调度，是一个接近最优的解决方案，但也是时间效率。这表明，所提出的AFCN是一个实用和有效的方法来解决供应链计划和调度问题，在完全分布式环境。
Supply chain planning and scheduling problems require manufacturers (project managers) to determine product configurations, select suppliers (contractors) for task allocation, and schedule project tasks while considering various production constraints among manufacturers and suppliers. Most relevant studies have focused on finding an optimal solution based on complete information provided by each enterprise in a supply chain. However, the practical implementation of complete information sharing is difficult, if not impossible, because of the fully distributed nature of the supply chain process. This work proposes an agent-based fuzzy constraint-directed negotiation (AFCN) model to solve problems associated with supply chain planning and scheduling, which are modeled as a set of fuzzy constraint satisfaction problems (FCSPs) that are interlinkedviainter-agent constraints. To accommodate the perspectives and interests of each enterprise in a supply chain, conflicts among FCSPs are resolved using the AFCN protocol through the iterative exchange of offers/counter-offers with limited information sharing and without privacy breaches. A proposed offer/counter-offer represents not only a set of acceptable solutions and preferences for an operational task but also the possibility of conflict in this area. For each FCSP, the incremental process of offer/counter-offer evaluation eliminates redundant and infeasible solutions. The sharing of limited non-strategic sensitive information among agents enables them to elucidate their opponents’ intentions through iterative negotiations, such that the agents can reach an agreement while ensuring that the solution to a project planning and scheduling problem is satisfactory. The AFCN model is also sufficiently flexible to incorporate different negotiation strategies such as competitive, win-win, and collaborative strategies, for various production environments. Herein, a numerical study was conducted to examine the practical viability and effectiveness of the proposed AFCN model. The experimental results show that the proposed AFCN model not only can generate a schedule that is comparable to a near-optimal solution but also is time-efficient. This indicates that the proposed AFCN is a practical and effective method for solving supply chain planning and scheduling problems in fully distributed environments.