LLM-based multi-agent simulations are increasingly adopted across application domains, but remain difficult to scale due to GPU memory pressure. Each agent maintains private GPU-resident states, including models, prefix caches, and adapters, which quickly exhaust device memory as the agent count grows. We identify two key properties of these workloads: sparse agent activation and an estimable agent invocation order. Based on an analysis of representative workload classes, we introduce invocation distance, a unified abstraction that estimates the relative order in which agents will issue future LLM requests. Leveraging this abstraction, we present ScaleSim, a memory-efficient LLM serving system for large-scale multi-agent simulations. ScaleSim enables proactive prefetching and priority-based eviction, supports diverse agent-specific memory through a modular interface, and achieves up to 1.74x speedup over SGLang on simulation benchmarks.
@article{arxiv.2601.21473,
title = {ScaleSim: Serving Large-Scale Multi-Agent Simulation with Invocation Distance-Based Memory Management},
author = {Zaifeng Pan and Yipeng Shen and Zhengding Hu and Zhuang Wang and Aninda Manocha and Zheng Wang and Zhongkai Yu and Yue Guan and Yufei Ding},
journal= {arXiv preprint arXiv:2601.21473},
year = {2026}
}