Research

Paper

TESTING March 24, 2026

ABSTRAL: Automatic Design of Multi-Agent Systems Through Iterative Refinement and Topology Optimization

Authors

Weijia Song, Jiashu Yue, Zhe Pang

Abstract

How should multi-agent systems be designed, and can that design knowledge be captured in a form that is inspectable, revisable, and transferable? We introduce ABSTRAL, a framework that treats MAS architecture as an evolving natural-language document, an artifact refined through contrastive trace analysis. Three findings emerge. First, we provide a precise measurement of the multi-agent coordination tax: under fixed turn budgets, ensembles achieve only 26% turn efficiency, with 66% of tasks exhausting the limit, yet still improve over single-agent baselines by discovering parallelizable task decompositions. Second, design knowledge encoded in documents transfers: topology reasoning and role templates learned on one domain provide a head start on new domains, with transferred seeds matching coldstart iteration 3 performance in a single iteration. Third, contrastive trace analysis discovers specialist roles absent from any initial design, a capability no prior system demonstrates. On SOPBench (134 bank tasks, deterministic oracle), ABSTRAL reaches 70% validation / 65.96% test pass rate with a GPT-4o backbone. We release the converged documents as inspectable design rationale.

Metadata

arXiv ID: 2603.22791
Provider: ARXIV
Primary Category: cs.AI
Published: 2026-03-24
Fetched: 2026-03-25 06:02

Related papers

Raw Data (Debug)
{
  "raw_xml": "<entry>\n    <id>http://arxiv.org/abs/2603.22791v1</id>\n    <title>ABSTRAL: Automatic Design of Multi-Agent Systems Through Iterative Refinement and Topology Optimization</title>\n    <updated>2026-03-24T04:28:07Z</updated>\n    <link href='https://arxiv.org/abs/2603.22791v1' rel='alternate' type='text/html'/>\n    <link href='https://arxiv.org/pdf/2603.22791v1' rel='related' title='pdf' type='application/pdf'/>\n    <summary>How should multi-agent systems be designed, and can that design knowledge be captured in a form that is inspectable, revisable, and transferable? We introduce ABSTRAL, a framework that treats MAS architecture as an evolving natural-language document, an artifact refined through contrastive trace analysis. Three findings emerge. First, we provide a precise measurement of the multi-agent coordination tax: under fixed turn budgets, ensembles achieve only 26% turn efficiency, with 66% of tasks exhausting the limit, yet still improve over single-agent baselines by discovering parallelizable task decompositions. Second, design knowledge encoded in documents transfers: topology reasoning and role templates learned on one domain provide a head start on new domains, with transferred seeds matching coldstart iteration 3 performance in a single iteration. Third, contrastive trace analysis discovers specialist roles absent from any initial design, a capability no prior system demonstrates. On SOPBench (134 bank tasks, deterministic oracle), ABSTRAL reaches 70% validation / 65.96% test pass rate with a GPT-4o backbone. We release the converged documents as inspectable design rationale.</summary>\n    <category scheme='http://arxiv.org/schemas/atom' term='cs.AI'/>\n    <published>2026-03-24T04:28:07Z</published>\n    <arxiv:primary_category term='cs.AI'/>\n    <author>\n      <name>Weijia Song</name>\n    </author>\n    <author>\n      <name>Jiashu Yue</name>\n    </author>\n    <author>\n      <name>Zhe Pang</name>\n    </author>\n  </entry>"
}