Learning Heterogeneous Resource-Constrained Task Allocation Using Concurrent Multi-Task Bandits

Task allocation is a critical aspect of multi-robot coordination, enabling the completion of complex tasks that would be intractable for individual robots. However, existing approaches to task allocation often assume that task requirements or reward functions are known and explicitly specified by the user in advance. In this thesis, we explore the challenge of forming effective coalitions for a given heterogeneous multi-robot team when task reward functions are unknown. To tackle this challenge, we first formulate a new class of problems, dubbed COncurrent Constrained Online optimization of Allocation (COCOA). The COCOA problem requires online optimization of coalitions in such a way that the unknown rewards of all the tasks are simultaneously maximized using a given multi-robot team with constrained resources. To address the COCOA problem, we propose an online optimization algorithm called Concurrent Multi-Task Adaptive Bandits (CMTAB), which leverages and builds upon continuum-armed bandit algorithms. Our experiments, which involve detailed numerical simulations and a simulated emergency response task, demonstrate that CMTAB is effective at balancing exploration and exploitation to efficiently optimize unknown task rewards while respecting the team's resource constraints. Our results suggest that CMTAB has the potential to enable effective task allocation for multi-robot teams in real-world scenarios, even when task reward functions are unknown.

Level thesis:degree_level

Masters

Department dc:contributor.department

Electrical and Computer Engineering

Grantor dc:publisher

Georgia Institute of Technology

Year dc:date.issued

2023

Author dc:creator

• Singh, Sukriti

Advisor dc:contributor.advisor

• Ravichandar, Harish

Committee members dc:contributor.committeemember

• Chernova, Sonia
• Gombolay, Matthew
• Hutchinson, Seth

• Task allocation
• Coalition formation
• Heterogeneous teams
• multi-arm bandits
• Active learning

Language dc:language.iso

en_US

Handle dc:identifier.uri

https://hdl.handle.net/1853/75155

OAI identifier oai:identifier

oai:repository.gatech.edu:1853/75155