GitLab

Detailed Description

When scheduling flights, the flight demand could also take into account the intervals of this flight to other flights on the same route operated by the airline.

A possible implementation would be calculation at scheduling. When scheduling a particular flight, the demand of the flight will decrease depending on the number of flights by the airline around the same time period (suggestion is to put 1 hour).

For instance, if 2 flights, which we call Flight 1 and Flight 3, flying on LAX-SFO by the same airline departs at 1100. Assuming Flight 1 will be scheduled first, they will take demand based on the current formula. Flight 3, which is scheduled later, will then have its passenger demand decreased by the number of passengers taken by Flight 1. This will effectively consider Flight 1 and 3 as the same flight, by demand calculations.

When flights have a slightly larger gap (eg. 5 minutes vs 35 minutes), but both intervals are still within a set period, the amount of passenger demand affected by other similar flights decrease. Using the same situation in the example above, Flight 3 will have 100% of Flight 1's passenger count deducted from its passenger demand, since they have an interval of 0 minutes. But if Flight 3 is moved to depart 30 mins after Flight 1, then a smaller portion (say, 40%) of Flight 1's passenger count will be deducted ftom Flight 3's passenger demand.

Context

The purpose of this is to slow down the growth of large, mammoth airlines, by preventing the spam of flight schedules which would be unrealistic in the real world context. An airline who launched 5 flights over a span of 15 minutes on the same route is unlikely to gain the same passenger count as compared to having these 5 routes spread out evenly across the day. This is in line with what the demand formula intends to achieve - which was to calculate demand per flight, not per route. So flights flown close to each other should therefore be calculated as though they are the same flight. In reality, airlines who fly on routes with such high demand would be likely to spread them out with at least 45 mins to 1 hour interval at the minimum, and even then, those routes are likely to be flown with 1 widebody per flight and not multiple narrowbodies.

To bring another realism concern into this suggestion, airlines typically find banked hubs more expensive because of the number of ground staff and gates required. Putting a whole bunch of flights together means they would need to lease more gates from the airport (which could be more costly, as compared to leasing only 1 gate but for a longer period of time) and also requiring more ground staff (to process the huge amounts of passengers checking in, and also the huge number of planes at the gates). While it may be difficult to actually simulate such costs, it would be simpler to start by clamping down on unrealistic flight schedules like these and putting the additional costs on these "uneven" flights.

Another reason is because of the ease of route creation which feeds into larger airlines' hands. Larger airlines are able to send multiple planes flying on the same route, simply by coping a created iternary multiple times. With this change, there would be additional effort required in creating routes, and the small change could make a huge difference in reducing "spam" on certain routes.