Efficient and Fair Decentralized Task Allocation Algorithms for Autonomous Vehicles
Abstract
This project produced four papers: 1) "Achieving Envy-freeness and Equitability with Monetary Transfers" addresses one of the original aims of allocating tasks and giving payments to ensure that all parties feel that the outcome was fair. Fairness can incentivize different groups of robots to cooperate; 2) "Multi-Robot Task Allocation-Complexity and Approximation" develops approximation algorithms to process the most number of tasks when there lower caps on robots requires for finish a task. It achieves a key goal of the research project of designing new algorithms for important robot-task allocation problems; 3) "A General Approximation Algorithm for Ordered Allocations" develops mathematical insights on how well CBBA type algorithms can approximate maximum welfare under submodular domains even under complex feasibility constraints. The result is central to our understanding of the welfare approximations achieved by CBBA-type algorithms. We focus on a class of algorithms that is more general than CBBA; 4) "Auction-based Allocation of Location-Specific Tasks" develops understanding of decentralized auction-based methods where the goal is minimize total distance travelled.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 21, 2022
- Accession Number
- AD1176789
Entities
People
- Haris Aziz
Organizations
- University of New South Wales