Scheduling a Distributed Permutation Flowshop with Uniform Machines and Release Dates

Global manufacturing optimizes production efficiency and cost, enabling enterprises to compete effectively. Distributed production is emerging for the trend of globalization and the requirements of diversified manufacturing, for which distributed scheduling plays an important role in enterprises enhancing efficiency and conserving energy. This study investigates a heterogeneous distributed permutation flowshop scheduling problem to minimize the makespan, in which release date and factory speed are incorporated to mirror a real-world production scenario. A mixed integer programming model is established to address this NP-complete problem using a business optimizer. The findings assist in identifying optimal properties for algorithm development and assessing the performance of a proposed branch and bound algorithm, which includes a problem-specific pruning rule and defined lower and upper bounds to reduce the solution space. For industrial scheduling, a dispatching rule named Dynamic Largest Processing Volume on the Fastest Machine First is proposed, offering asymptotic optimality and ensuring uninterrupted production. Additionally, a discrete artificial bee colony algorithm with a novel availability rule and an abandonment criterion is introduced to improve efficiency. Simulation results demonstrate the efficacy of the proposed algorithms. Note to Practitioners —Distributed production leverages information technology to integrate geographically decentralized manufacturing units for cost reduction and production enhancement. This mode finds extensive application in automobile, electronics, and medical equipment manufacturing. In mass production scenarios, the online Dynamic Largest Processing Volume on the Fastest Machine First rule provides a viable alternative to optimal algorithms, delivering a convergent schedule rapidly and preventing unnecessary delays. In complex industrial settings, the discrete artificial bee colony algorithm offers effective solutions without relying on a mathematical model, proving more efficient in such environments.