University Professor Lavanya Marla, along with Dartmouth Professor Vikrant Vaze and MIT Professor Cynthia Barnhat, developed a new aircraft scheduling model that can reduce the negative impacts of unexpected disruptions on flight schedules.

According to the U.S. Bureau of Transportation Statistics, aircraft delay has been a serious problem for airline companies. The cascading effects of one flight delaying could lead to a series of flight delays or even cancellations along the following schedules. These seemingly inevitable delay problems have caused billions of dollars lost each year in the airline industry.

According to Marla, the new models are more applicable to the aircraft scheduling context than the existing ones. It allows the airline manager to better analyze the situation and grasp more controls over the system than before.

“Under the huge context of airline network, people haven’t applied to cases where there are millions of variables, millions of possible decisions,” Marla said. “The conclusions of the paper indicate the combination of art and science involved in such modeling paradigms and show that the new approaches we propose can be better than existing approaches if intelligently applied.”

As the leader of the study, Marla said the research process presents a lot of challenges.

“This research is probably the first of its kind,” she said. “The comparison of different methodologies didn’t exist and it is really tricky for us to understand how these methods compare to each other in practice.“

The research, which started in 2014, includes analyzing past data and comparing data between different existing models. By repeating this process, the research team was able to understand how the existing methods control the airline system and develop new models that cover numerous variables of real-life circumstances.

The new paradigms do not merely apply to aircraft scheduling. Marla said the models are applicable to a wide range of network-based resource allocation problems subject to uncertainty, such as those occurring in train or bus scheduling, freight or goods transportation and even problems such as Unmanned Aerial Vehicles routing.

“I am always looking for capable graduate students who have a strong mathematical background and good programming skills to join my research group,” Marla said. “I have also worked with strong undergraduate students and they have learned a lot about real data and research in the process of assisting graduate students.”

[email protected]