How Southwest Airlines created schedule reliability against disruptions

"Developing SkySYM to simulate the operation of our highly complex, point-to-point route network was no simple task,” Southwest Airlines Senior Manager of Operational Performance Jeff Borges said. “With SkySYM, we will better understand the impact that different network designs have on our operational performance. We are very pleased with the dedication, expertise and commitment that Optym displayed at every step of this journey, and with the results we are seeing."

Jeff Borges Southwest Airlines Senior Manager of Operational Performance

Background and challenge: 

If airline schedules are optimized solely for profitability, it may negatively impact a schedule’s reliability and operational metrics, which are difficult to measure before flying the schedule. Southwest Airlines, one of the largest airlines in the world, creates flight schedules months in advance. It wanted to develop a software solution to predict schedule reliability and operational KPIs with high accuracy during the planning phase. Southwest also needed to identify and adjust critical flights that affect overall reliability and on-time performance in the network. 

Solution: 

Since airline operations are prone to irregularities and disruptions, it can expose a planned flight schedule to operational risks and uncertainties. To address this problem, we developed SkySYM, a simulation-guided optimization system, to assess the on-time performance of a flight schedule before it’s flown and to improve its reliability by making schedule changes, without impacting profitability. 

We worked closely with Southwest’s network planning and network operations departments to understand and model its business operations. We built a digital twin of Southwest’s operations using discrete event simulation and modeled all important operations, such as flight delays, maintenance disruptions, weather, crew and passenger movements, airport congestion, and schedule recovery of deviations and disruptions. 

Using advanced machine learning, we modeled airline processes and validated them against historical data, achieving 98% accuracy for critical KPIs. We developed new techniques to identify problem areas, including flights that cause critical delays in the network. We adjusted flight times and aircraft rotations to improve schedule reliability without compromising profitability. SkySYM combines our simulation and optimization algorithms with an advanced decision support system, where users can define data inputs, prepare scenarios, run simulations and view schedules using intuitive visualization tools. 

Results: 

SkySYM has become the standard system used by Southwest’s network planning and network operations departments to analyze and improve schedules for operational performance before publishing them. Southwest uses SkySYM’s predictive analytics and optimization capabilities to understand a schedule’s operational performance and potential bottlenecks and to improve schedule reliability. Our studies showed that SkySYM could improve on-time performance by up to 5% without impacting profitability, saving the airline tens of millions of dollars in operational costs annually. 

See for yourself