A Cooperative Coevolution Algorithm with a Heterogeneous-Island-Based Hyper-Heuristic for Cloud-Edge-Device Collaborative Scheduling
In distributed manufacturing, cloud servers, edge servers, and shop-floor devices are coordinated to jointly execute computational services and production operations. Scheduling such systems requires three tightly coupled decisions: task offloading, operation sequencing, and device assignment. This work develops CCHIHH, a structure-aligned framework that decomposes the decision space into semantically coherent blocks, assigns each block an independent adaptive learner, maintains parallel search trajectories through heterogeneous islands, and safeguards productive convergence with a stability gate. Experiments on instances ranging from 50 to 500 tasks with up to 800 devices show the advantage of block-specific adaptation, stronger performance against evolutionary baselines, and slower performance deterioration under resource degradation and communication inflation.