Distribution System

Feature Roadmap
,
1

What?

The heart of the ASTD: Instead of generating travel requests per airport “in bulk” once a day, the Distribution System uses Individual Travel Request Generation (point-2-point) and does an individual search per request. It also doesn’t generate those requests from airport to airport, but from actual “locations”…roughly 150k towns and cities around the world.

Why?

There are several main reasons for this:

  1. Demand distribution becomes more natural, as bookings trickle in over time.
  2. Demand is not airport-based anymore. Instead, airports are merely gateways, which is very important to shape the advantages and disadvantages of remote, low-cost airports vs. large hubs.
  3. Demand becomes a bit more dynamic, as airports don’t use a fixed set of destinations that get booked every day, but locations are pseudo-randomly picked from a large set, causing natural fluctuations.

When?

Whether we will be able to transplant this into current-gen AirlineSim is somewhat open. We pulled a lot of tricks to make it work in ASTD and we might just not get it to work in the existing system, but we’d certainly love to try.

This obviously depends on the whole “ASTD feature set”, so Individual Travel Request Generation (bulk), Custom Booking Classes, Fares (base feature) and Advanced Interlining.

But it also paves the way for super-advances stuff like return flight bookings and much more advances fare rules.

2

Can you expand more on the ‘remote, low-cost airports vs. large hubs’? Is this more like a distance-based model or something else? Because if it’s a distance-based model, then you will need to also differentiate international v.s. domestic airport so that you don’t have a domestic or domestic focus airport, which is usually closer to the city centre but has much fewer slots, being developed into a mainly international hub and one or few airline taking all the advantages of the airport proximity.

For example, RJTT is always the preferred airport for most travellers to Tokyo, but because of this, the international slot allocation is very limited and highly restricted, which, in general, pushes demand to RJAA or even to RJNA when RJAA was slot-restricted back in the old days. If there are no international slot restrictions modelled at RJTT, you will very likely have players, especially international market players, filling in all the slots at RJTT first (and it probably has this demand in real-life), which might end up with a route map that is highly disproportional and unrealistic nor fair for any players who actually want to set up a JP airline.

There are also airport route restrictions in general, e.g., RJNA, which is extremely limited in slots, disallows any commercial flights other than some FDA regional operations, while all the commercial flights are pretty much at RJGG, even though RJNA is literally at the city centre and would be the preferable airport by demand.

Just curious how this feature could be achieved because it seems like to simulate the demand properly, there will be huge research involved and multiple flags and limitations being applied, which probably also need to be displayed in-game so that players know what route scheduling is actually legal or not without having all the background knowledge.

3

Going to address this first to clear up a potential misunderstanding: AirlineSim is not a model train set where the goal is to mimmick reality as closely as possible. If that where the case, I could just import real-world schedule data, start the simulation and everyone has “fun” watching what happens.

The goal is to have a plausible simulation given the parameters we feed in. Will they always be “realistic”? No of course not. In fact, they might be completely and utterly removed from reality (just have a look at our current maintenance system). This also means that anything that’s an edge-case in reality likely won’t work just like that in AS. Purely domestic airports being a prime example as long as we don’t feed that information into the simulation. That doesn’t make it “unfair”, though. It just means it’s different than reality.

So concerning the actual question: At the end of the day, it’s a matter of airport infrastructure. To many passengers, flying from HHN will actually make a lot more sense than to fly from FRA. But because HHN is, in comparison, a tiny single-runway airport, it will never be able to facilitate the sort of “international hub operations” that make FRA what it is.

This means that for anything that one can operate from HHN, HHN won’t necessarily be at a disadvantage. It’s just going to end up being more direct traffic that doesn’t rely on network effects as much.

At some point, other attributes might also play a role, like the general attractiveness of the airport (tbd what this would even mean…) or how well it is connected to public transport. And of course whether an airport is “international” or not. We actually have that flag in the database for many airports already, but is doesn’t have any effect yet. It’s technically beyond the DS as a base feature…but might warrant a new entry on the feature roadmap :slight_smile:

1 Like
4

completely agree. Not asking to make it exactly like real-life or most of my current operations is going to be shutdown :wink:

That’s the main part of my concern. Because the AS is made to make everything possible, players could make the HHN an ‘international hub operation’, and if HHN is set to be preferable to FRA because distance-wise, you will have HHN at an advantage compared to FRA, and anyone who can take up all the slots HHN slots at the start, even though being unrealistic destinations and everyone else suffers and can only use slot at the less preferable FRA. The current system doesn’t have that issue because you still have solid, if not higher, demand at FRA, which means FRA, which has more slots, will be ideal to fly to with the slots to support it. The current system actually has some correlation between total slots and demand, which I am very afraid would go away under this change and making the slot issues far worse.

I’m not saying direct enforcement for domestic only vs international, but I think to make the slot issue addressed, I think something has to be done so that everyone doesn’t need to fight for demand for smaller airport closer to the city and early birds get all the preferable slots and demand, while the larger airport outside the city being the leftover for everyone else.

Edit: Just look at the location of HHN and I’m not sure about what you mean by

because it looks far from everywhere.

I didn’t know Frankfurt enough and thought HHN was more of a city airport, which seems to be not the case. Maybe I just do not understand the methodology the proposed system is going to be. i understand and completely agree there is definitely a need to make demand expand further away from the existing airports when slots are limited at this airport; what comes into my mind would be like RJAH, KLGB, etc, but all I want to point out is there is definitely a need of avoiding city airports, such as EGLC, RJOO, RJTT, etc, having over- or preferable demand than their slots allow, even though considering all the other factors, such as infrastructure and location etc., they clearly should be. Maybe I just do not understand the purpose of this feature at all and what I brought out is not relevant to this feature.

5

What I mean was: In the ASTD model, our database contains any place in the real world that has a population of 500 people or more, roughly 170k locations in total. This means that there are going to be a lot of locations that are a lot closer to a third- or fourth-tier airport like HHN than to a real-world mega hub like FRA. In those locations, flying from HHN will generally be preferable to traveling to FRA first and then fly from there.

But given the infrastructure constraints of HHN, you won’t be able to operate a hub there. You’ll be able to run some direct flights to popular destinations and you might be able to do so very cheaply due to the small airport size (which is exactly what the likes of Ryanair are doing), but you won’t be able to offer international connections etc. because even if transfers are technically possible at a smaller airport, you won’t have the capacity for proper hub waves.

For city airports, the situation is similar, but also different. An airport like London City will have massive demand but obviously it’s utterly impossible to operate a large network from there. What can and will happen is that other network carriers will fly into LCY to suck passengers into their hubs in other locations, which will be very attractive.

A gap in this model is that LCY likely charges ridiculous fees…something that we don’t model in AS just yet.

Is it a bit clearer now?

2 Likes
6

It’s clearer, and it sounds good, but i think there will be huge issues with slots based on my current understanding.

The current HHN has 5 slots per 5 minutes. You can definitely run a hub there, and people do. I run hubs with waves from many airports with similar sizes/demands myself. So I think it’s a very bad assumption based on the current game slot model. It sounds like you actually need to make HHN transfers impossible, which is probably not good either because you are going to limit the number of airports that can have transfers significantly.

This probably would be required to make this model work, so the player also has to charge a very high price to make a profit and force any low-fare pax not to use the airport. I think you might even need to set the minimum price for all the routes to be very high plus max slot allocation per airline to prevent slot blocking. With the current game model, e.g., slots, pricing etc, it will be pretty reasonable to assume players will run profit-losing flights at the beginning to just get all the slots they could get at these preferred airports.

I can see the reason for the other ASTD feature dependency now, but I think there would be also quite some work to make sure it won’t make the slot issue even worse.

7

I am not sure I can follow. All the things you describe are already true today. The only thing that’s different is the source of demand. And in cases like LCY, not even that is all that different due to ground networks.

But to stick with the example of FRA vs. HHN: Due its remoteness, HHN would likely still have overall lower demand than FRA. But slots are exactly the same. I guess, you could build a “hub” in HHN as it is. But at 60 take-offs and landings per hour, it would never be a particularly large one.

Slot blocking has always been and will always be an issue. Finding measures against it is it’s very own subject and not something that, in my opinion, directly affects the DS.

8

I don’t believe HHN would be an issue. City airports like LCY would be because, if I understand it correctly, this increases the demand at LCY, which is a smaller airport with far fewer slots. Based on my current understanding, DS will make it let’s say 10 bars due to it’s location, instead of the current 6 bar. That would be a nightmare for slots because people will try to schedule as many flights to it as LHR, which has 10 slots per 5 mins instead of 3. The slots will run out much faster due to excess demand compared to right now. Basically, you have demand at the airport that it would not run out even if you take out all the slots, and based on the airport runway length, you have to run smaller aircraft in the first place.

9

Yes. And what about that is “wrong”? :slight_smile:

Or, asking in a different way: What would you feel would be “right” in such an (egde-)case?

10

This is wrong because not all cities have the luxury of building multiple large airports within the city, and the DS will make it reverse demand. I’m talking about ITM, HND, DCA, NKM etc, and they are heavily slot restricted or slot regulated in real-life, and generally don’t have slots to support the actual demand, which in the game currently shown as low demand, while the alternatives them, KIX, NRT, IAD, NGO, etc, shown as high demand in game even though their actual demand is lower compared to the city airports. This is actually very common for cities that have city airports, not an ‘edge case’. And certain countries, like the US, don’t build ground infrastructure to support the non-city airports either. so you have to actually force demand to go up because there is no ground network to support that. YMX failed to replace YUL for that very good reason when you have no slot restrictions at YUL and no ground network at YMX.

I don’t have A prefect solution in my mind, because in real-life there isn’t one either. I think there is always some kind of slot allocation in play though. If you ask me to come up a solution right now, my current solution will be making them domestic flights only, and plus route length and frequency limitation for countries like US. That will make sure they can have the demand boost from the DS, while not making the slot issue worse off because of the extra demand

notes: HND has a high demand in game also, but it doesn’t have as much issue as it has a lot of slot by default, but the issue is still there, because the game currently kind of equally weighs Tokyo demand at HND and NRT, while realistic speaking (maybe DS) it weights demand far higher at HND than NRT, which will still end up with slot issue

The HHN and FRA discussion is actually a separate topic, because that’s more like an LLC airport vs a traditional hub, which DS probably won’t do much harm if not benift only

11

All the restrictions you mention boil down to the number of slots we have in the game. We could apply more real-world-inspired restrictions, and I haven’t ruled those out. But at the end of the day, either we misunderstand each other or we simply disagree :sweat_smile:

Yes, due to its location, HND might be more attractive than NRT. Which means players are going to prefer it. Until slots run out. Which is when NRT becomes more attractive merely due to the availability of slots. I do not see a fundamental issue with this.

As soon as capacity at HND can’t cover demand, pax will look for alternatives and might chose flights through NRT. That’s part of the appeal of the new system.

Whether and how we add further flavor and differentiating attributes to airports is a different topic, but one that’s literally part of the opening post:

The “shaping” aspect here might be a tad too vague. But in principle, it covers all the things you mention.

1 Like
12

I think this comes down to whether you consider this as helping the older player consolidate even more because they get the more preferable airports from the start, and can just hold it forever. Personally, I think it will make it even harder for new players to enter the market, which I thought is one of the issues for the current game.

I just want to point out if some issues are considered, not pushing for how you implement it. After a decade on AS, I think I learnt to just accept how the system is. I’m not sure if you will ever implement something like this for the older servers, but I really hope we don’t have route restriction implementations again where there is a lot of work for players to change while not actually solving the slot issues at all. I rather reschedule for actually achieving something :wink:

13

But Martin’s logic (which actually reflect real world), doesn’t mean there wouldn’t be space for new players, but same as in real world, the game will get more interesting, because there will be more opportunity for niche markets - yes, at some point, the big players will squeeze out the competition from the frequented airports, but from what I understand, there will be much more opportunity then today to work different models in the game, so there might be room for a “Ryanair type” ULCC airline, that will serve more remote airports at a considerably lower price point, or a regional airline that will focus on small airports alone. I am actually excited to see how well will the new demand and booking class model work.

5 Likes
14

Yes, and?

Isn’t that exactly the real situation in quite a lot of city airports? (DCA, LCY comes to mind instantly)

They ARE slot blocked in real life. You will never have a perfectly fair and equal distribution of slots in AS. Players who join a server when it’s started will always have a major advantage, but I don’t see anything that’s wrong about it.

IMO the bigger issue is slot wasting. Players who operate a million regional jets or small narrow bodies on trunk routes. I know that this is also done by the US3 in real life to some extent.

But what would tremendously in my eyes would be a restructuring of the cost structure. At the moment it’s way too cheap to operate a regional jet compared to medium-sized narrow body. The scale effects need to be larger.

1 Like
15

Actually this is not how real life works. DCA has route length restrictions, and it is also a domestic airport, so that’s not the same as current AS already, where you can have ALL the demand at airports no matter if it’s international or domestic. These restrictions curve the airport usage and demand in real-life, which you don’t have it in AS. The basic idea of DS is great, but applying it to the current AS slot/airport system, in my eyes, would be a disaster.

It’s true that airlines hold slots in real life, but it’s also the case that if you don’t use them properly, the airport can remove the slots from you, which can’t be done in AS. Delta famously lost its international slot at RJTT due to improper / lack of usage. For busy airports, you very likely need to apply for the slot first, which isn’t doable in AS. For AS, even right now, what you usually end up with is one or two airlines taking up all the slots at these slot-limited airports, which doesn’t happen realistically. If Delta gets all the slot allocations at DCA, the airport authority is going to get into trouble very fast. I was just suggesting if there is no change to the current slot system, the DS might make this even more likely (which might not be expressed properly in the previous posts). In real life, it’s actually not common for you to end up with a true monopoly in one market because 2 out of the US 3 are not going to be out of business one day, and even so, there are airlines which already exist, can fastly add in capacity to avoid monopoly. For AS, it’s not the case. If you have one of the airlines in the market leaving the server, usually the remaining one will actually take up all the demand, which would be a true monopoly. I have seen it happen so many times on Kaitak, and it’s also what I did when FENCC01 left my core market. There are fortress hubs, but even so, I don’t think it would be that common for you to see one airline taking 90% of the slot at the airport.

The trunk routes are the largest issue, as I have also stated over time, but it’s not the only one. I run multiple wave hubs, and I actually used up slots very fast, even though I’m willing to upgrade to widebodies to save slots instead of maximising profits. Think about what happens when you make the whole city demand available at DCA and also make it the most preferable airport in the city without any route or slot limitations; you are going to make the airline dominate the airport very fast and, in some sense, dominate the whole city demand for both domestic and international routes. Basically, you are making one legacy carrier in one city while any newcomers are LCCs because, basically, you are making both IAD and BWI the further away LLC option. Realistically speaking, I don’t think it’s as common as you think it is, nor is it good for the server. What should happen is when they fill the DCA to a certain extent, they have to start their international hub journey at IAD. They still get the early bird advantage at DCA, a preferred airport in the city, as in real life, and they will be forced to develop more routes and a new hub at IAD, a less preferred airport in the city, as in real life.

This is going off-topic. I am actually always against the fare structure change for RJ or narrowbodies because that’s how you build your network in AS and also in real life. Airlines always start small with props/RJ, maybe some narrowbodies, so applying a higher landing fee makes no sense at all. This only makes any new airline in the market harder to grow. How to solve the slot issues, in my opinion, is still limiting the route frequency. If one player uses one prop per day on one route and another player runs 100 320s flights on the same route, the answer is always that the second player is more likely to be considered slot blocking, not the first one. Actually, the route frequency limits can be considered something like a slot allocation solution because, basically, you disallow any large percentage of slots used by one or few particular routes. You don’t need to limit how many props/RJ are being used by players; you need to force the players to upgrade to widebodies when the route demand is reached (of course, while being feasible. The current widebodies turnaround makes it hard if you care about profit margin).

When Martin said this, it pretty much addressed my concern because the plan is not just applying the DS to the current slot/schedule system directly, which was my main issue.

16

Question:
Are rushers/swifters, passengers who want to arrive at their destination as soon as possible and are time sensitive, planned in any way, shape or form?
If so, then not only more airlines can coexist in the same area, but also will incentivise airlines being time-sensitive oriented, meaning that their strategy is to bring passengers to their destination as fast as possible, as well as making the usage of fast aircrafts more viable

17

We can define an arbitrary amount of “passenger profiles” and time sensitivity (or even preferred times-in-the-day/days-of-the-week) is definitely part of each such profile. We won’t be able to fully exploit this in the first iteration (bulk request generation) but if and when we are able to implement point-2-point request generation and return flights, these aspects will play a major role.

1 Like