Green Computing Resource Allocation and Task Scheduling Algorithm Wins ORSC 2024 Operations Application Award

On October 21, the 17th Annual Conference of the Chinese Operations Research Society (ORSC 2024) was held in Guiyang, where the joint project "Resource Allocation and Task Scheduling Algorithms under Green Computing" by Professor Zhang Guochuan’s team from Zhejiang University and the Ant Group Optimization Intelligence team received the ORSC Operations Application Award.

ORSC, organized by the Chinese Operations Research Society, aims to strengthen connections among researchers, share the latest theoretical, algorithmic, and application advances in operations research, and promote international collaboration and research hotspots.

The Operations Application Award, established in 2002, recognizes outstanding applications of operations research that contribute to social and economic development.

Operations research is a highly applied interdisciplinary field used in transportation, finance, resource development, urban planning, production scheduling, industrial engineering, market analysis, and military domains, enhancing decision‑making and management effectiveness.

The awarded project, originating from the CCF‑Ant Green Computing Fund, aligns with Ant Group’s 2030 net‑zero goal. It abstracts the global scheduling optimization of Ant’s large‑scale computing clusters into three interlinked stages, applying operations‑research modeling and solution techniques at each stage. This enables optimal resource planning across the full lifecycle (configuration, deployment, migration), saving tens of thousands of CPU cores, reducing thousands of tons of CO₂, and extending to large‑model training/inference clusters to improve GPU utilization.

Related research results have been published at top conferences such as STOC‑2024 and VLDB‑2024. Relevant papers include:

· A Nearly Quadratic‑Time FPTAS for Knapsack: https://dl.acm.org/doi/abs/10.1145/3618260.3649730

· OptScaler: A Collaborative Framework for Robust Autoscaling in the Cloud: https://www.vldb.org/pvldb/vol17/p4090-lu.pdf