The ticket-price-routing problem is not new, even if it is interesting. There was an article on HN a few months ago, IIRC, talking about how ITA uses tens of th

The ticket-price-routing problem is not new, even if it is interesting. There was an article on HN a few months ago, IIRC, talking about how ITA uses tens of thousands of lines of Common Lisp, with hand-tuned assembly, to get their optimization queries to finish within their tight space (given the huge set of possible combinations) and time bounds. My point was that the presentation's slides are very cursory in their treatment.