Partial Replanning for Decentralized Dynamic Task Allocation

TL;DR

This paper introduces CBBA-PR, a decentralized algorithm for multi-UAV teams that efficiently reallocates new tasks during missions, balancing rapid response with coordination quality.

Contribution

It extends the CBBA algorithm with partial replanning, enabling fast, conflict-free task reallocation without full reoptimization, reducing communication and runtime.

Findings

CBBA-PR achieves faster convergence in simulations.

It maintains conflict-free solutions during online task addition.

Reduced team size further improves efficiency.

Abstract

In time-sensitive and dynamic missions, multi-UAV teams must respond quickly to new information and objectives. This paper presents a dynamic decentralized task allocation algorithm for allocating new tasks that appear online during the solving of the task allocation problem. Our algorithm extends the Consensus-Based Bundle Algorithm (CBBA), a decentralized task allocation algorithm, allowing for the fast allocation of new tasks without a full reallocation of existing tasks. CBBA with Partial Replanning (CBBA-PR) enables the team to trade-off between convergence time and increased coordination by resetting a portion of their previous allocation at every round of bidding on tasks. By resetting the last tasks allocated by each agent, we are able to ensure the convergence of the team to a conflict-free solution. CBBA-PR can be further improved by reducing the team size involved in the replanning, further reducing the communication burden of the team and runtime of CBBA-PR. Finally, we validate the faster convergence and improved solution quality of CBBA-PR in multi-UAV simulations.