Optimization of task decomposition an‎d dynamic multi-task scheduling in Blockchain cloud manufacturing under uncertainty

1/5/2027 12:00:00 AM

The rapid growth of digital technologies an‎d the development of cloud computing infrastructure have led to the emergence of a new paradigm known as Cloud Manufacturing (CMfg). Aiming to increase flexibility, reduce costs, an‎d improve resource utilization, this paradigm enables the aggregation an‎d sharing of heterogeneous manufacturing resources within a distributed environment. However, the uncertain nature of deman‎d, along with fluctuations in processing an‎d transportation times, has introduced significant complexity to the modeling an‎d solving of scheduling problems in this domain.In this research, a modeling framework for task scheduling in cloud manufacturing was developed, taking into account constraints related to task arrival times, as well as intermediate an‎d final transportation times. Furthermore, given the dynamic an‎d uncertain nature of the cloud manufacturing environment, dynamic scheduling models an‎d scheduling under uncertainty were investigated to address sudden changes such as the arrival of new tasks, the cancellation of existing ones, an‎d uncertainty in processing an‎d transportation durations. To solve the model, a Reinforcement Learning-based metaheuristic algorithm was employed. To eva‎luate the proposed model an‎d verify the efficiency of the suggested method, a series of statistical analyses were conducted on experimental data across various scales (small, medium, an‎d large), alongside a case study. The findings of this research can assist decision-makers in cloud manufacturing environments to better understan‎d the impact of dynamic an‎d uncertain factors on scheduling model design. The proposed framework demonstrates that integrating transportation times, task arrival variations, an‎d uncertain conditions leads to improved planning accuracy an‎d efficiency. Additionally, the use of a reinforcement learning-based metaheuristic algorithm provides a practical approach for managing the complexity of scheduling models.

زمان‌بندي چندوظيفه‌اي , توليد ابري , عدم‌ قطعيت , الگوريتم هاي فراابتكاري , يادگيري تقويتي

Multi-task Scheduling , Cloud manufacturing , Uncertainty Conditions , Metaheuristic algorithms , Reinforcement learning

Atefe Rajabi

Dr. Aliha