Air Force Research Laboratory, or AFRL, will host the second iteration of the Airlift Challenge competition Oct. 30, 2023, through Feb. 19, 2024. To identify new solutions, AFRL is inviting international researchers and university students to register to apply their skills and develop unprecedented new algorithms that achieve on-time deliveries while improving efficiencies.
The online event, which is open to the public through challenge.gov, seeks to advance state-of-the-art planning algorithms for executing airlift operations. The Challenge winners will be presented in June 2024 at the International Conference on Automated Planning and Scheduling (ICAPS) in Banff, Alberta, Canada.
The first Airlift Challenge was held online from Jan. 23 through Feb. 27, 2023. There were two winners, and the results were presented at the International Society for Optics and Photonics, or SPIE. A paper was also published with writeups from each winner on the SPIE website.
"A large demand and tight deadlines make airlift operations difficult to plan even under ideal conditions," said Dr. Andre Beckus, AFRL machine learning researcher at the Information Directorate "Unexpected disruptions only further complicate the problem, potentially introducing major delays and stressing planning software to its limits."
During the competition, participants can submit agents for immediate scoring to see their rank on a real-time leaderboard. In the end, three overall winners will be announced. Their proposed solutions will be open-source.
"The Airlift Challenge provides a simulation environment in which artificial intelligence, or AI, agents can interact," Beckus said. "Agents will be tested against a series of scenarios of ever-increasing complexity with evaluation ending when a time limit is reached or when the agent misses too many deliveries."
Teams, researchers and university students are encouraged to register and try out the starter kit ahead of the competition start date. For questions, please contact here.
Artificial Intelligence Analysis
Objective:
The objective of the Airlift Challenge is to identify new solutions to advance state-of-the-art planning algorithms for executing airlift operations, using AI-based agents.
Current State-of-the-Art and its Limitations:
The current state-of-the-art techniques for executing airlift operations are difficult to plan even in ideal conditions, and unexpected disruptions complicate the problem, potentially introducing delays.
Whats New:
The Airlift Challenge provides a simulation environment in which AI agents can interact. Agents will be tested against a series of scenarios of ever increasing complexity.
Target Audience and Impact:
The challenge is open to international researchers and university students, and will be hosted by the Air Force Research Laboratory. The proposed solutions will be open-source, and if successful, could lead to improved efficiency in airlift operations.
Risks Involved:
The risks involved in pursuing this approach include the potential for missed deliveries, delays, and other unforeseen complications.
Cost :
The cost of pursuing this approach is not specified.
Timeline:
The second iteration of the Airlift Challenge will take place from October 30, 2023 to February 19, 202
4. Winners will be announced in June 2024.Success Metrics:
Mid-term and final success metrics include the number of successful deliveries, the time taken to complete the deliveries, and the efficiency of the proposed solutions.
Score: 8/10
DARPA is likely to be interested in this challenge as it provides an opportunity for international researchers and students to test their skills and develop new algorithms to achieve on-time deliveries while improving efficiencies. The proposed solutions will be open-source if successful, which makes it a viable option for DARPA. The risks involved in pursuing this approach are acknowledged, but no cost and timeline estimates are provided, which limits the ability to assess the viability of the project.
