This research focuses on the development of recommendation algorithms for heterogeneous computing power platforms. By analyzing the task details submitted by users, the system can efficiently and accurately identify the optimal computing resource allocation plan, enabling personalized and intelligent recommendations. The aim is to maximize computing power utilization, reduce both user costs and time expenses, and enhance the overall performance and efficiency of data center operations and networks. The research is organized around four key areas: (1) Multi-dimensional data collection and preprocessing for user tasks and computing resources; (2) Task and resource modeling for computing power service recommendations; (3) Research on heterogeneous computing resource recommendation techniques based on user computational needs; (4) Performance evaluation methods and optimization strategies for computing power service recommendations.

Collaborating Institution: Hunan University