The encyclopedia · Engineering & Operations · Operational decision · 2005–2007
Delta's optimizer picks which alliance flights to sell under its own code
Choosing codeshare flights used to take Delta staff one to two days; an optimizer picks the revenue-maximizing set in hours, worth up to $50M a year.
Delta Air Lines
the move
Delta sells seats on flights its alliance partners operate by codesharing. Which flights it codes heavily changes its revenue, and the choice is constrained by alliance rules, government rules and union agreements.
The old approach was manual: staff spent one to two days choosing codeshare flights by hand, with no optimization and no reliable view of total network effect.
The change was to automate selection as a revenue-maximizing optimization over the candidate set of codeshare flights, subject to all those rules.
The system raised Delta's potential operating revenue by up to $50 million a year and cut the planning cycle from days to a few hours.
why it works
- Each codeshare choice reallocates revenue across many flights, so local picking is wrong.
- Alliance, government and union constraints make the feasible set hard to find by hand.
- Automating the search yields a near-optimal set, not merely an acceptable one.
- Speed matters because codeshare volume and rule complexity kept growing.
what transfers
A revenue decision that looks local actually ripples across the network. Optimize the set together and against your real constraints, not flight by flight.
what came after
The codeshare profitability system compressed a days-long manual judgment into a few hours of optimization, delivering a Delta-estimated up to $50 million a year in extra operating revenue and a far shorter planning cycle.
references
- Development of a codeshare flight-profitability system at Delta Air Lines
- Development of a Codeshare Flight-Profitability System at Delta Air Lines
spotted an error? The archive wants to know.