MSSE Thesis Defense: Tekevwe Akor, "Improving Decentralized Task Allocation For Multiagent Systems"
Friday, April 16, 2021
MSSE Thesis Defense
A Method For Improving Decentralized Task Allocation For Multiagent Systems In Low-Communication Environments
Advisor - Professor Jeffrey Herrmann
Committee member – Professor Shapour Azarm
Committee member – Assistant Professor Michael Otte
Task allocation is an important aspect of coordination in multiagent systems. However, low communication restricts the information exchange needed for task allocation. As a result, a lot of decentralized task allocation algorithms perform worse in low-communication environments.
In this thesis, I present a method that optimizes communication between agents, using the CBAA algorithm, in low-communication environments. The method determines when an agent should communicate and tries to predict the information that will be received from other agents, thus maximizing the value of each communication.
This method is compared to other decentralized task allocation algorithms such as CBAA, CBBA, PIA, DH and HIPC at different levels of communication in a ship protection scenario. The communication model used is the Rayleigh-Fading model. Results show that this method performs comparably to the above algorithms and reduces the amount of communication among agents.