Scheduling rules learned for clearer and wider use in systems
Factorized Scheduling Principle: Learning Interpretable and Transferable Policies via Structured Additive Functions
Machine LearningArtificial Intelligence
Summary
Scheduling means picking one item from many based on certain rules. The authors propose a way to learn simple, understandable rules for scheduling that can work well even when conditions change. Their method breaks down these rules into smaller parts that consider items alone and in pairs, making the results easy to explain. Tests show it works well on made-up and real problems and can handle bigger or different tasks without retraining.
What this means in practice
- •For manufacturing schedulers: Create clear and adaptable rules to prioritize jobs efficiently without manual tuning in changing production environments.
- •For cloud resource managers: Implement easy-to-understand policies that prioritize tasks and scale well when allocating computing resources automatically.
Authors
Hong Je-Gal, Hyun-Suk Lee
Abstract
Scheduling problems arise from repeatedly selecting one item from a set of candidates based on their states. These problems often reduce to assigning priority scores and choosing the highest-ranked item. In this work, we propose a factorized scheduling principle (FSP) framework to learn interpretable and transferable scheduling rules. The FSP framework represents system states as condition distributions and decomposes a global scheduling principle into additive univariate and pairwise components with identifiability constraints. The scheduling principle enables the framework to maintain a simple priority-based structure during deployment. This principle is learned by using a policy-based objective combined with a temporal-difference signal defined on the condition distribution. Experiments on synthetic and realistic scheduling tasks demonstrate the FSP framework's strong performance, interpretability, and zero-shot generalization across different system scales.