Auction-Consensus Algorithms for Decentralized Task Allocation in Multi-Robot Systems

Author

• Jose Alberto Rodriguez, The University of Texas Rio Grande Valley

Date of Award

8-1-2026

Document Type

Thesis

Degree Name

Master of Science in Engineering (MSE)

Department

Electrical Engineering

First Advisor

Wenjie Dong

Second Advisor

Qi Lu

Third Advisor

Rogelio Soto

Abstract

This thesis investigates methods for improving decentralized multi-robot task allocation (MRTA) through enhancements to auction-consensus coordination. The first contribution of this work introduces the Auction-Consensus Algorithm with Loss Mechanism (ACALM), a deterministic extension that incorporates loss-aware bidding. Through dynamic loss propagation, agents rebid in a manner that better reflects global coordination impact under the Single-Task Robot, Single-Robot Task, Instantaneous Assignment (ST-SR-IA) assumptions. The second contribution replaces handcrafted bidding rules with learned neural bidding policies trained under a centralized training with decentralized execution framework. Using reinforcement learning, agents learn cooperative bidding strategies that improve allocation quality under the Multi-Task Robot, Single-Robot Task, Instantaneous Assignment (MT-SR-IA) assumption. Together, these results demonstrate that both structured loss mechanisms and data-driven bidding adaptation significantly narrow the performance gap between decentralized auction-consensus algorithms and centralized optimal assignments.

Comments

Copyright 2026 Jose Alberto Rodriguez. All Rights Reserved. https://proquest.com/docview/3409153901

Recommended Citation

Rodriguez, J. A. (2026). Auction-Consensus Algorithms for Decentralized Task Allocation in Multi-Robot Systems [Master's thesis, The University of Texas Rio Grande Valley]. ScholarWorks @ UTRGV. https://scholarworks.utrgv.edu/etd/2000