Task Structure and Scalability in Multi-Robot Authoring

Christian Marowski; Byung-Cheol Min

cs.ROOct 15, 2026

Task Structure and Scalability in Multi-Robot Authoring

Authors: Christian Marowski, Byung-Cheol Min

Organizations: SMART Lab, Luddy School of Informatics, Computing, and Engineering, Indiana University Bloomington, Bloomington, IN, USA

Abstract

Multi-robot authoring can reduce completion time by distributing drawing work, but the resulting speedup depends on task structure and the allocation-routing pipeline. We investigate these relationships by representing drawings as geometric stroke graphs and comparing component-level (single-token) allocation with shared-token (decomposed) allocation, which permits multiple robots to work within a connected component. A simulation study of 385 drawings and teams of one to eight robots shows diminishing speedup gains under single-token allocation and larger mean speedups under the shared-token pipeline, each normalized by its own single-robot completion time. Shared-token execution also exhibits lower idle-time shares at larger team sizes, although it does not reduce completion time for every drawing. Comparisons across structural attribute groups reveal associations between structural descriptors and differences in scaling behavior, with greater separation between several attribute groups at larger team sizes. These findings characterize empirical scaling in the evaluated pipelines and highlight the importance of considering within-component workload and graph structure alongside component count. Additional materials are available at https://sites.google.com/view/multirobotauthoring/home.

Figures & tables

Fig. 1 : Task structure and allocation granularity on a Quick, Draw! drawing (“bus”). (a) Input vector strokes. (b) Stroke graph with inserted crossing vertices (red) and connected components. (c) Single-token allocation assigns each component to one robot. (d) Shared-token allocation splits work inside a component across robots. Bars show schematic execution timelines.

Fig. 2 : System overview. Vector strokes form a stroke graph. Single-token allocation treats each component as indivisible, whereas shared-token allocation permits work within a component to be distributed among robots (Algorithm 1 ). The reactive layer executes the routes using priority assignment and anchor-based yielding (Algorithm 2 ).

Setting	Value/status
Drawing corpus	385 graphs
Robot counts	$1,2,3,4,6,8$
Robot model	Homogeneous discs; center draws
Depot	$(0,0)$
Timestep, speed, disc radius	$0.02$ s; $80$ units/s; $2.25$ units
Conflict handling	Reactive priority/yield protocol

TABLE I : Simulator settings and experimental configurations.

Fig. 3 : Physical multi-robot authoring setup (right) with three HamsterS robots (left) collaboratively executing the selected airplane drawing.

Fig. 4 : Mean per-drawing speedup relative to each pipeline’s own single-robot completion time. Token and Decomposed denote single-token and shared-token allocation, respectively.

Fig. 5 : Paired relative completion-time differences

A(G,n)

across drawings. Green markers show medians; positive values favor shared-token allocation, negative values favor single-token allocation, and the dashed line indicates equal completion times.

Fig. 6 : Robot-time composition under single-token allocation. Each segment is the mean of the corresponding per-drawing robot-time fractions, expressed as a percentage. Each stacked bar sums to 100%.

Fig. 7 : Robot-time composition under shared-token (Decomposed) allocation, using the same per-drawing normalization and averaging as Fig. 6 .

Fig. 8 : Mean within-pipeline speedup difference,

\Delta S(G,n)

, across low, medium, and high structural attribute groups.

Method	Robots	Simulation	Physical	Difference
Token	2	1.71 $\times$	1.63 $\times$	$-4.8\%$
Token	3	1.71 $\times$	1.59 $\times$	$-7.3\%$
Decomposed	2	1.76 $\times$	1.69 $\times$	$-3.7\%$
Decomposed	3	2.31 $\times$	2.10 $\times$	$-9.1\%$

TABLE II : Comparison of simulated and physical speedup

Multi-Robot Systems

Multi-robot systems coordinate teams of robots to plan, perceive, communicate, and act together in shared tasks and environments.

See all 205 papers

Scaling Inference-Time Compute with Teams of Coding Agents
Oct 12, 2026Keenan Samway, Daniel D'souza, Omar Darwiche Domingues +3Inference-Time Scaling Inference-Time Compute Allocation
Cohere Labs · Cohere
OSDAG: Online Scheduling for Efficient Multi-Robot Collaboration
Jun 13, 2026Thanh Nguyen Canh, Thang Tran Viet, Phuc Van Dinh +2Multi-Robot Systems Large Language Model-Based Robot Planning
School of Information Science, Japan Advanced Institute of Science and Technology, Nomi, 923-1211, Ishikawa, Japan. · University of Engineering and Technology, Vietnam National University, 10000, Hanoi, Vietnam. · Department of Robotics, Hanyang University, Ansan, 15588, Gyeonggi, Korea.
FORMICA: Decision-Focused Learning for Communication-Free Multi-Robot Task Allocation
Feb 20, 2026Antonio Lopez, Jack Muirhead, Carlo PinciroliMulti-Robot Task Allocation Multi-Agent Resource Allocation
Dept. of Robotics Engineering, Worcester Polytechnic Institute, Worcester MA 01609, USA

Oct 12, 2026J/K move · Enter open · S save

Scaling Inference-Time Compute with Teams of Coding Agents

Keenan Samway, Daniel D'souza, Omar Darwiche Domingues, Marzieh Fadaee+2

Cohere Labs · Cohere

Inference-Time Scaling Inference-Time Compute Allocation Multi-Agent Coding Multi-Agent Resource Allocation Software Engineering Agents Cost-Aware Inference

Open paper AskPDF arXiv

Multi-agent systems promise a new axis for scaling inference: rather than having one agent think longer or sample more attempts, a task is split across many agents that work in parallel and coordinate. However, in practice, more agents can easily produce duplicate work, coordination failures, and higher costs. We therefore formulate team design as a compute allocation problem: starting from a solo agent, where should additional inference compute go to reach the reward-cost Pareto frontier? We build a minimal multi-agent harness from the ground up and trace this frontier by varying, one at a time, the workflow stage at which a lead delegates to its workers (planning, implementation, or verification), the models that form the team, and the number of workers. On ProgramBench, teams outperform solo agents but cost three to seven times as much. With the right compute allocation, however, a heterogeneous 17-agent team comes within 1.4 points of a strong homogeneous 5-agent team at 70% lower cost. This efficiency stems from three findings. First, planning is the most valuable use of workers: in a homogeneous team whose lead decides how many workers to spawn, a lead that delegates only planning to its workers, and then implements and verifies the code itself, matches the reward of a lead that delegates to workers throughout planning, implementation, and verification, at half the cost. Second, cheap models pay off as planners but not as implementers: a strong lead with cheap planners recovers 47.5% of the reward gain of an all-strong team for only 13.9% of its additional cost, whereas cheap implementers lower reward. Third, planners keep improving reward as the team grows, whereas implementers and verifiers plateau beyond four workers. Together, our results offer a guide to spending an agent budget where it counts, given the task, cost, and latency constraints a user has.

Abstract

Multi-Robot Systems

Explore similar work

Scaling Inference-Time Compute with Teams of Coding Agents

OSDAG: Online Scheduling for Efficient Multi-Robot Collaboration

FORMICA: Decision-Focused Learning for Communication-Free Multi-Robot Task Allocation