Skip to content
Preprint

MARS: A Monte Carlo Tree Search-based Adaptive and Responsive Scheduler

Aug 2026 · 0 citations · 30 references
Computer Science

TL;DR

This work introduces MARS (Monte Carlo Tree Search-based Adaptive and Responsive Scheduler), a training-free HPC scheduler whose optimization goal is configurable through a reward function rather than baked into a learned model.

Abstract

Modern High Performance Computing systems depend on static heuristics and manual administration for job scheduling and reservation management. Deep Reinforcement Learning (DRL) has shown promising scheduling performance but requires historical training data and fixes the optimization goal at training time, forcing operators to retrain whenever priorities shift. We introduce MARS (Monte Carlo Tree Search-based Adaptive and Responsive Scheduler), a training-free HPC scheduler whose optimization goal is configurable through a reward function rather than baked into a learned model. MARS uses a lightweight discrete-event simulator to explore the future consequences of scheduling decisions within a strict time budget, adapting to the configured reward at each scheduling cycle. We evaluate MARS on year-long production workloads from two systems at Argonne Leadership Computing Facility -- 4,360-node Theta and 560-node Polaris---under two reward functions: wait-time minimization (MARS-CW) and utilization maximization (MARS-CU). Unlike DRL and heuristics, which only react to the current queue or wait for backfill to find holes, MARS exploits look-ahead to proactively drain the system and plan around future reservations, packing the system to avoid the fragmentation and utilization drop that typically precede reservation windows. MARS-CW reduces tail wait time by 64% on Theta and 43% on Polaris over the production WFP heuristic, while MARS-CU recovers utilization in the 48 hours leading into maintenance, demonstrating that MARS can target either objective via reward reconfiguration.

View source

Similar papers

#reinforcement learning Open access Sep 2026

Research on Multi-Objective Task Scheduling Optimization and Reinforcement Learning Decision-Making Methods for Edge-Cloud Collaborative Scalable Information Systems

INTRODUCTION: Edge-cloud schedulers must coordinate latency, energy, load balance, and deadline compliance under changing demand while keeping task-arrival and throughput units physically consistent.OBJECTIVE: This study evaluates MORL-ECSO under an auditable, paired-seed simulation protocol and compares it with tuned...

Li-Na Guo, Cheng-Yu Sun · 0 citations
Book Open access Aug 2026

EDA Job Scheduling Using Reinforcement Learning with Adaptive Macro Actions

Adaptive Job Selection (AJS), a reinforcement learning-based agent that learns to schedule pending jobs for reduced job waiting and completion time on LSF clusters for EDA workloads, is presented, becoming the first open-source, deployable RL-based scheduler designed for production EDA environments.

Yiming Shao, Aijun An, Michael Spriggs et al. · 0 citations
#reinforcement learning Open access Dec 2026

Optimizing Resource Allocation in Cloud Computing Environments using Reinforcement Learning

A deep reinforcement learning framework for intelligent cloud resource allocation that jointly optimizes resource utilization, Service Level Agreement compliance, infrastructure cost, and energy efficiency, and adapts to workload distribution shifts within 200 episodes without manual retuning is presented.

Msr Prasad · 0 citations
Preprint Aug 2026

A Smallest-Need-First Job Scheduling Framework with Adaptive Optimization of Idle Node Counts for Energy-Efficient HPC Systems

SNF-ICON is presented, an event-driven controller combining smallest-need-first (SNF) gang scheduling, predictive wake timing, and adaptive warm-spare control that reduces average waiting time relative to the FCFS-based baseline in all six cases and remains close to at least one heuristic energy baseline in five.

Reza Pulungan, Raka Satya Prasasta, Santana Yuda Pradata et al. · 0 citations
Preprint Aug 2026

BOOSTEDSOSA: Accelerated Inferencing for Low Variance Stochastic Online Scheduling

BOOSTEDSOSA is introduced, a dual-FPGA ML-assisted Scheduling architecture that integrates a Machine Learning predictor for expected processing times, with a novel temporal-aware training policy, enabling its use in existing HPC systems.

Adam H. Ross, Riccardo Revalor, Aryan Singh et al. · 0 citations

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.