# Rack Position Optimization in Large-Scale Heterogeneous Data Centers

arXiv: https://arxiv.org/abs/2504.00277 · HTML: https://mlsanity.com/paper/2504.00277/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 7 of its 22 references that are in the map.
- First posted on arXiv 31 Mar 2025.

## Builds on (most cited first, 7 of 7)

- [PyTorch: An Imperative Style, High-Performance Deep Learning Library](https://mlsanity.com/paper/1912.01703/) (2019)
- [Attention, Learn to Solve Routing Problems!](https://mlsanity.com/paper/1803.08475/) (2018)
- [POMO: Policy Optimization with Multiple Optima for Reinforcement Learning](https://mlsanity.com/paper/2010.16011/) (2020)
- [Combinatorial Optimization with Policy Adaptation using Latent Space Search](https://mlsanity.com/paper/2311.13569/) (2023)
- [Learning-based Two-tiered Online Optimization of Region-wide Datacenter Resource Allocation](https://mlsanity.com/paper/2306.17054/) (2023)
- [RackSched: A Microsecond-Scale Scheduler for Rack-Scale Computers (Technical Report)](https://mlsanity.com/paper/2010.05969/) (2020)
- [Energy Efficient Placement of Workloads in Composable Data Center Networks](https://mlsanity.com/paper/2101.08244/) (2021)

## Abstract (arXiv)

As rapidly growing AI computational demands accelerate the need for new hardware installation and maintenance, this work explores optimal data center resource management by balancing operational efficiency with fault tolerance through strategic rack positioning considering diverse resources and locations. Traditional mixed-integer programming (MIP) approaches often struggle with scalability, while heuristic methods may result in significant sub-optimality. To address these issues, this paper presents a novel two-tier optimization framework using a high-level deep reinforcement learning (DRL) model to guide a low-level gradient-based heuristic for local search. The high-level DRL agent employs Leader Reward for optimal rack type ordering, and the low-level heuristic efficiently maps racks to positions, minimizing movement counts and ensuring fault-tolerant resource distribution. This approach allows scalability to over 100,000 positions and 100 rack types. Our method outperformed the gradient-based heuristic by 7% on average and the MIP solver by over 30% in objective value. It achieved a 100% success rate versus MIP's 97.5% (within a 20-minute limit), completing in just 2 minutes compared to MIP's 1630 minutes (i.e., almost 4 orders of magnitude improvement). Unlike the MIP solver, which showed performance variability under time constraints and high penalties, our algorithm consistently delivered stable, efficient results - an essential feature for large-scale data center management.