Skip to the content.

Swissport - Innosuisse funding to apply reinforcement learning with human feedback to staff dispatching

In collaboration with Swissport we successfully received funding from Innosuisse for an innovative project on tailored deep reinforcement learning methods with human feedback (RLHF) to enhance resource-efficient dispatching and operational delay reduction. The co-learning approach aims to foster trust in AI systems and maintain a high degree of human control while simultaneously improving system performance.

Centered Image

The innovation project will enable Swissport to further improve their operations and thereby solidify their position as a market-leading ground handling agent through better predictions of upcoming turnaround delays and the mitigation of deviations from normal operations with optimal resource allocation. The AI expert system, based on reinforcement learning, will support human dispatchers in their daily work and reduce workloads during normal operations. During periods of disruption, a successful human-AI teaming and co-learning approach will leverage the strength of the human operator and the AI algorithms. Employee feedback is therefore critical for the reinforcement learning algorithm to improve its solution quality and suggest solutions to the human dispatcher that are closer to what a colleague will do than what a computer will come up with. Overall, the solution will help improve the quality of service provided to Swissport’s customers and keep Swissport at the forefront of aircraft ground handling.


Back


Written on August 11, 2023 (last modified: August 11, 2023)
comments powered by Disqus