讲座题目：Artificial intelligence models in airline scheduling and operations: What is the future?

主讲人：吴政隆教授（澳大利亚新南威尔士大学）

讲座时间：2023.6.7（星期三）上午9:00

讲座地点：一教404教室

讲座对象：全校师生

主要内容：Reinforcement Learning-based AI models are shown successful in many fields. The application of RL-based AI models in the airline business is still limited. This talk presents a RL prototype to solve the aircraft line maintenance scheduling problem (LMSP). LMSP is concerned with scheduling a set of maintenance tasks with service deadlines during an aircraft’s ground time. To address this problem, we introduced a heuristic-based scheduling model and solved LMSP with a RL-based algorithm. The experimental results based on Qantas data demonstrated that the proposed AI-enhanced scheduling solutions achieved near-optimal schedules for all test instances. Although the developed AI model was trained by specific schedule instances, the deployed AI model demonstrated near-optimal scheduling results across a set of different test instances with varying sizes. Another focus of this talk is on the comparison between AI models and OR models and how optimisation-based methods can co-exist with AI models in the future.

主讲人简介：

Dr. Wu is currently an Associate Professor at UNSW Aviation who specialises in airline operations management, scheduling, airport terminal planning, airport retail development, passenger choice behaviour, and airline big data analytics. Many of his past projects helped industry partners save millions of dollars in operating costs or enhance product sales and revenues. Dr. Wu leads a team of researchers who have specialised skills to deliver value for industry partners through world-class aviation research at UNSW Sydney. UNSW Sydney is ranked 45 among universities in the world. Dr. Wu has spent time both in the public and private sectors in transport and aviation. Wu’s recent research projects include a fuel policy optimisation project (funded by Qantas), an aircraft line maintenance scheduling project (Qantas), AI models for air passenger ticket booking (Virgin Australia, Velocity Frequent Flyer), and airport planning and retail simulation projects. Dr. Wu’s papers have been published in top journals including Transportation Science, Computers and OR, J. of Air Transport Management, Transport Research Part A, C & E, Tourism Management, and J. of Travel Research. Wu’s book on airline operations and delay management has sold more than a thousand copies worldwide since 2010. Wu’s research outputs have been adopted by aviation insurance companies, consulting & investment firms, airlines, ground handling companies, and airports. A recent application was Wu’s output in 2014 on airline alliance terminal co-location by the Chicago O’Hare Airport in 2016. Wu’s publication in 2015 on the airline delay data coding system was adapted by IATA in 2018 as a new delay data collection scheme for the airline industry. IATA is currently evaluating Wu’s output in 2020 on improving airport terminal design standards.