Crew Optimization for Presence on Scheduled Airlines


Scheduled crew resources with no legality violations


A U.S. government agency responsible for staffing crew on passenger airlines sought to develop a next-generation crew scheduling system. We developed a single-pass optimization system that scheduled 100 percent of the available crew resources, while improving distribution of crew on flights and honoring lifestyle requirements.

Business Problem

Our client was a large U.S. government agency with crew assigned on passenger airline flights. It needed to optimize its schedule to cover the highest priority flights with certain probability. The legacy tools involved cumbersome workarounds and manual intervention, yet they assigned according to a duty roster less than 50 percent of the required resources. The legacy process took the better part of a month to complete. The agency’s objective was to develop a tool that could optimize the crew schedules in a single run and roster close to 100 percent of the available resources. The optimized schedule would have to adhere to strict engagement roles governing daily flight and duty times. Further, the schedule would have to honor contractual lifestyle rules, along with leave and ground duty assignments. The complexity of the problem was compounded by a flight schedule many times larger than any single large airline.

Our Approach

We created a team of business and IT experts who had deep domain expertise and prior experience in developing software solutions for the airline industry. First, we studied the operating environment and validated business requirements and assumptions provided to us against current manually created rosters. Next, we formulated an integrated approach, including mixed integer programming (MIP) and heuristics that the user could run to create optimized rosters for all crew members.

The team developed a user interface to run the model and view the KPIs and roster output graphically. The system also created data output to use in Excel for detailed analysis. At the conclusion of the engagement, we worked closely with the agency to assess the study’s outcomes.

Key Benefits

  1. The optimizer rostered 100 percent of the crew from multiple bases with no legality violations.
  2. The entire end-to-end optimization ran in a fraction of time compared to just a single step in the legacy process, which took the better part of a month.
  3. Our single-pass optimization process replaced multiple separate runs demanded by the legacy system.


We created a model that was generic and predictable to find the optimal solution in a wide range of parameter settings. We engineered the optimization process so the client could invoke it during operations to provide recovery to the initial schedule when assumptions were updated.