Delwedd Flaen

Efficient Scheduling for GPU-Based Neural Network Training via Hybrid Reinforcement Learning and Metaheuristic Optimization

Wedi'i Gadw mewn:
Cyhoeddwyd yn:Big Data and Cognitive Computing vol. 9, no. 11 (2025), p. 284-325
Prif Awdur: Du, Nana
Awduron Eraill: Wu, Chase, Hou Aiqin, Nie Weike, Song Ruiqi
Cyhoeddwyd:
MDPI AG
Pynciau:
Scheduling
Performance measurement
Neural networks
Integer programming
Costs
Graph theory
Optimization
Decision making
Workload
Algorithms
Quality of service
Machine learning
Heuristic
Completion time
Workloads
Heterogeneity
Heuristic methods
Run time (computers)
Mynediad Ar-lein:Citation/Abstract
Full Text + Graphics
Full Text - PDF
Tagiau: Ychwanegu Tag
Dim Tagiau, Byddwch y cyntaf i dagio'r cofnod hwn!
Crynodeb:On GPU-based clusters, the training workloads of machine learning (ML) models, particularly neural networks (NNs), are often structured as Directed Acyclic Graphs (DAGs) and typically deployed for parallel execution across heterogeneous GPU resources. Efficient scheduling of these workloads is crucial for optimizing performance metrics such as execution time, under various constraints including GPU heterogeneity, network capacity, and data dependencies. DAG-structured ML workload scheduling could be modeled as a Nonlinear Integer Program (NIP) problem, and is shown to be NP-complete. By leveraging a positive correlation between Scheduling Plan Distance (SPD) and Finish Time Gap (FTG) identified through an empirical study, we propose to develop a Running Time Gap Strategy for scheduling based on Whale Optimization Algorithm (WOA) and Reinforcement Learning, referred to as WORL-RTGS. The proposed method integrates the global search capabilities of WOA with the adaptive decision-making of Double Deep Q-Networks (DDQN). Particularly, we derive a novel function to generate effective scheduling plans using DDQN, enhancing adaptability to complex DAG structures. Comprehensive evaluations on practical ML workload traces collected from Alibaba on simulated GPU-enabled platforms demonstrate that WORL-RTGS significantly improves WOA’s stability for DAG-structured ML workload scheduling and reduces completion time by up to 66.56% compared with five state-of-the-art scheduling algorithms.
ISSN:2504-2289
DOI:10.3390/bdcc9110284
Ffynhonnell:Advanced Technologies & Aerospace Database