Cloud manufacturing task decomposition method considering resource compatibility and competitiveness

To address the issue of disconnection between task decomposition and resource matching in cloud manufacturing, this paper proposes a task optimization decomposition method based on the HDBSCAN clustering algorithm. First, the task decomposition process is clarified, according to the task decomposition principles, the overall task is decomposed into indivisible meta-tasks. Secondly, a model of subtasks-resources-capabilities is constructed, transforming the ‘many-to-many’ mapping relationships into ‘one-to-one’ mapping relationships. Subsequently, considering task-resource compatibility, competitiveness, and the adaptive adjustment requirement of task granularity with the platform resource quantity, a hybrid approach combining genetic algorithm and HDBSCAN clustering algorithm is proposed to solve the model, resulting in the final cloud manufacturing task decomposition. Finally, the feasibility and effectiveness of the algorithm are verified by means of case studies. This method considers both task-resource compatibility and competition among resources, achieving optimization of task decomposition in cloud manufacturing, bridges the gap between task decomposition and resource allocation.

针对云制造（cloud manufacturing）中任务分解与资源匹配脱节的问题，本文提出一种基于HDBSCAN聚类算法（HDBSCAN clustering algorithm）的任务优化分解方法。首先，明确任务分解流程，依据任务分解的核心原则，将整体任务拆解为不可再分的元任务（meta-tasks）。其次，构建子任务-资源-能力关联模型，将原本的"多对多"映射关系转化为"一对一"的映射关系。随后，综合考量任务-资源兼容性、资源竞争力，以及任务粒度随平台资源规模的自适应调整需求，本文提出一种融合遗传算法与HDBSCAN聚类算法的混合求解方法，对所建模型进行求解，最终得到优化后的云制造任务分解方案。最后，通过案例研究验证了所提算法的可行性与有效性。该方法同时兼顾任务-资源兼容性与资源间的竞争特性，实现了云制造场景下的任务分解优化，有效弥合了任务分解与资源分配之间的脱节问题。