Connecting flights/transiting passenger

Started by TommyC81, November 08, 2009, 07:43:21 PM

TommyC81

Before getting started; Yes, I've searched and found a couple of topics about this:

https://www.airwaysim.com/forum/index.php/topic,12045.0.html
https://www.airwaysim.com/forum/index.php/topic,10670.0.html
https://www.airwaysim.com/forum/index.php/topic,3360.0.html
https://www.airwaysim.com/forum/index.php/topic,6081.0.html

etc. etc.

But I still feel the need to vent some thoughts and ideas, so bare with me. Here goes...


The current system is pretty straightforward, point-to-point routes only and very simplified competiton, a bit too simplified as there is a lot of talks now about anti-monopoly regulations etc. To go beyond having just a little fun and have a few planes and really be able to expand, you currently have to be based on a large high-demand airport (fixed to real-world numbers, not dynamic). Even if multilple bases would be added to the game, it still wouldn't really add anything except more simple hard-core expansion options, it wouldn't really add to "being smart". This takes a bit of the edge away. From what I understand the current issue is the staggering amount of calculations that would be necessary to have truly transiting passengers. But what about a simplified, expandable version of this.
Let's start off by some conditions to limit the amount of possibilities for a "transfer" (which can later be expanded/relaxed):
* No "second destinations" allowed in your schedules
* It can only be your own flights
* PAX accepts a maximum of 1 transfer
* PAX accepts a maximum transfer time of 3 hours
This should greatly limit the number of options for a transfer, and can of course be relaxed later on depeding on computational limits etc.


So how would this work?
Every time a route-pair is created/edited/removed the "connections table" (database table) is updated with consideration to the above mentioned rules.

Example:
You already have connection A - B - A, departing 08.00 and returning to base at 12.00.
You create connection A - C - A, departing 14.30 and returning to base at 18.00.
When creating second connection the following happens:
A connection "A - C" is created in the database with calculation information such as price, service levels, aircraft, type of aircraft, capacity, departure time, total flight time and other variables necessary to calculate PAX distribution. Also during the creation of connection "A - C" a check is also made to see if a transfer is possible according to the above mentioned rules.
As there is an arrival from a different destination by your own airline within 3 hours before the flights departure, a connection is possible between B - C via A!
A connection "B - C" is created in the database with calculation information necessary to calculate PAX distribution incl. but not limited to an average service level, via, total flighttime, total price etc.

When time to determine distribution of passengers from B to C, all flights are compared (the entered direct connections for those airports and the transfer connections) and depending on PAX preferences the PAX are distrubuted on the avialable choices between B and C.


This would of course require a few more calculations to determine exact capacities of the different legs used, and what is prioritised in terms of direct PAX vs. connecting PAX, and other calculations/presentations to make this work fully as intended.

In combination with adjustments to take departure time (high-demand times) more into account and adjusting flights to make transfers possible, this would add to the "be smart" instead of "be more" part of the game. Since it's obviously impossible to have all flights at all time to depart in the peak hours it would make it possible for smaller airlines to have clever routes with lots of connecting passengers between smaller destinations. As an example a small airline flying small, low-capacity airplanes with short turn-around times would be able to have connecting flights and compete with direct flights as total flying time would not be much more. And as smaller planes are cheaper to operate they should also be able to have pricing levels that make even connecting flights attractive vs. direct flights.

Along with this system I would also like it more clearly shown the prices and service levels of other airlines, in order to provide a possibility to make sensible choices in terms of pricing, service levels, departure times and type of aircraft. I've read the aeguments for hiding this information but disagree as it makes the game anonymous and almost turns it into a vs. computer game. All in all, I would definately ask for more "smart" than "more" improvements (read: connecting flights, better distribution of PAX vs. multiple bases). At least until the "smart" improvements are implemented.

An obvious extension of this is of course to reset the fixed demand figures between airport pairs and recalculate them depending on population, airport size (uptake area) and other nearby airports, and let natural hubs be created depending on where airlines set up their bases. A larger capacity airport (open 24H) would of course allow more flights then. And in an extension having demand be dynamic and based on most utilised aiport in the area, capacity etc. I've read the discussion about it, and maybe now is the time to recalculate the fixed numbers for demand and let natural hubs evolve?

Just some quick ideas before the nightshift. Thoughts?


samomuransky

Quote from: TommyC81 on November 08, 2009, 07:43:21 PM
* PAX accepts a maximum transfer time of 3 hours
Why?

Quote from: TommyC81 on November 08, 2009, 07:43:21 PM
* It can only be your own flights
Why? We should be able to make codeshare agreements.. This could make alliances more reasonable in game.

But I'm more for system of manual adding connection services.. For example, I have route A-B and A-C.. but system wouldn't sale B-A-C route automaticaly, but I would need to set it first manually..

Also, minumum connecting times - there should be some minimum time (based on airport size, maybe real number of flights in specific daytime). It would work exactly like turn-around times - if you just use minimum connection time, there would be high possibility of delay (due to waiting for connecting pax). Adding +30-40% of time would set "possibility of delay" to very low number.

powi

Because:

Quote from: TommyC81 on November 08, 2009, 07:43:21 PM
--- From what I understand the current issue is the staggering amount of calculations that would be necessary to have truly transiting passengers.---
--- This should greatly limit the number of options for a transfer, and can of course be relaxed later on depeding on computational limits etc. ---

TommyC81

#3
Quote from: Samo on November 09, 2009, 04:13:44 PM
Why?
Why? We should be able to make codeshare agreements.. This could make alliances more reasonable in game.

But I'm more for system of manual adding connection services.. For example, I have route A-B and A-C.. but system wouldn't sale B-A-C route automaticaly, but I would need to set it first manually..

Also, minumum connecting times - there should be some minimum time (based on airport size, maybe real number of flights in specific daytime). It would work exactly like turn-around times - if you just use minimum connection time, there would be high possibility of delay (due to waiting for connecting pax). Adding +30-40% of time would set "possibility of delay" to very low number.

Thanks Powi for clarifying reason why, maybe wasn't obvious  :)

Defining your own connecting flights may also be an option to limit the number of calculations necessary, that would also leave room for more precisely adjusting pricelevels, service while waiting, seats with priority for connecting passengers etc. Of course, some rules could still apply if a connecting flight is possible or not. Had not really thought of this option. Defining your own connecting flights would still require a lot of administration compared to a couple of settings and automation of this, but could possibly be worthwhile.

Either way, a better presentation of the competition is still desireable, like prices and servicelevels, in order to actually be able to make sensible choices. In comparison Airline Empires (another onlie airline game) does show pricing levels of other airlines and is overall much harder as soon as you're competing with others on different routes. Spamming a route with cheap capacity is extremely costly, while in Airwaysim, it's just a matter of small fuel cost and a monthly lease.

Runner

What do you think about this: not every flight and airport needs to be connected to every other flight or airport. There are only certain routes, I believe, on which there is a demand for connecting flight. You would have to come up with some kind of 'transisting passengers index', in order to really be able to simulate the connection thing.

Sami

Without reading all this. Would you prefer to determine the connection flights and prices yourself (ie. choose which connections are offered), or let the system be fully automated?  At least some price controls would be needed in my mind.

Riger

Quote from: sami on November 09, 2009, 07:40:21 PM
Without reading all this. Would you prefer to determine the connection flights and prices yourself (ie. choose which connections are offered), or let the system be fully automated?  At least some price controls would be needed in my mind.

IMHO

Passengers will naturally connect between airports, with or without the help of an airline.

For this reason, I'd automate the pricing system (as it is now) but leave the option for a player to tinker with the prices, offer special prices for pax that fly a-c via b when flown with that airline etc..

Best Regards
Richard


TommyC81

#7
Quote from: sami on November 09, 2009, 07:40:21 PM
Without reading all this. Would you prefer to determine the connection flights and prices yourself (ie. choose which connections are offered), or let the system be fully automated?  At least some price controls would be needed in my mind.

My initial thought regarding connections and pricing was an automated system for determining connections offered and a general setting for discount on the cheapest fare. It would make the system very easy but would perhaps be a bit restrictive in terms of price control. Maybe a good starting point?

A manual system that would show possible connecting flight offerings which can be accepted and then fine adjusting the prices and possibly also how many seats are reserved for your direct PAX vs connecting PAX could work, but would be a very tedious task unless very streamlined as soon as you open a new route which enable lots of connections

So, to answer your question; I'd prefer automated connection offerings and an improved pricing management system overall to compare competition and streamline the pricing process (incl. a research system to see what estimated effect on profit and loadfactors a price change would have, we have staff employed for this kind of research  :) ).

Quote from: Riger on November 09, 2009, 08:04:41 PM
IMHO

Passengers will naturally connect between airports, with or without the help of an airline.

For this reason, I'd automate the pricing system (as it is now) but leave the option for a player to tinker with the prices, offer special prices for pax that fly a-c via b when flown with that airline etc..

Best Regards
Richard



I wouldn't say the pricing system is automated, rather only suggesting. To keep comparing, there is an option for "Auto pricing" in Airline Empires which sort of tries to achieve maximum profit (or lack thereof if heavy competition), though not as good as you could by manually adjusting prices (there is a research option to try different prices and their outcome). But the current system which basically hides prices of the competition and their service levels make it barely a guessing game to compete (i.e. lower the price a couple of %, wait a couple of days and see if it made a difference). Airwaysim has a lot of strengths, in particular the presentation, the route planning system (which gives a feeling for actual peak times and achieving maximum use of your aircraft) among other things, and with a few more things added Airwaysim will be even more miles ahead.

Interesting reading: http://www.airlineempires.net/content/view/48/49/

ekaneti

I think we should start with one stop passenger first before we do connections. I have played other games with connections and it is very complicated. ie how do you control for connecting pax versus locals? What if you want only locals, what about 10% connects, 20% etc? Take that concept to a hub with 100 flights and it is impossible to do manually

powi

Mostly automatic, if possible (manual is better than nothing). However there should be some manual control of this, because some airlines could choose to not offer connecting flights (IRL Ryanair). Choosing to offer connecting flights would mean higher demand for customer service personnel etc and delayed/canceled flight should cost to the airline, because it need to re-route passengers, maybe to another carriers. Marketing costs should also be significantly higher (mostly for game balance).

Jona L.

Why are all of you writing books here?!
you could make millions (alltogether) if you would just write Novells! ;D ;D ;D

Well... I think you should either only let PAX connect to alliance intern routes, or to your own ones only!!

that is my opinion, but mabe you have better ideas ;);)!

ali5541

To add my coins to this great idea, I would have the system create 2 separate route pages. In other words, you would the normal route page where you can modify your current routes (the "Manage Routes" page) and the connecting flights page where the system automatically creates connecting flights. Using that page, you can do whatever changes you would like for connecting flights. It is a bit more organized this way and less confusing.
Member since July-2007

Yb

After some thinking I have an idea how to simplify the idea a bit.

a) All connection routes have to be created manualy (yes they have to be as in real world).

b) after you create a connecting route it will be a normal one (as per normal direct flight) but less pax will be willing to travel with you (depending on the waiting time, aircraft type, price).

By creating the routes by hand you are making sure that there will be no more calculations and even if there was still you use the normal chart - no extra numbers. After you create this you can then just mark which routes will be connected (i.e. you have a route 0031 from Amsterdam to Prague and then 0250 to Moscow - if you have at least hour in between these flights and you can mark them as a connecting flight.

Its a bit of a simplified concept but I think it would be a good start.

Branmuffin

Just thinking about the sheer number of possible connections that can be created even if you only have 70 or 80 unique routes, it is staggering to imagine having to create all those flights manually, one by one... (unless i am misunderstanding something about the manual creation suggestions?)

I don't have any hard knowledge of real world airline scheduling systems, but I have a hard time believing that when you book a flight online at an airline's website, it's anything other than a computer somewhere that is simply comparing your inputs of departure point, destination, and requested dates, with a database of all the airline's individual route legs to spit out all possible connections.

I don't see how anything other than at least a partially-automated system would work in AWS.  

As for pricing, I think the best solution is to first take the sum of the user-defined prices for each leg, and then factor in a discount that reduces this sum to a reasonable total fare for the whole flight (since, as I understand it, in real life you don't pay for each leg you fly.. you pay one price for the whole flight that is derived based on the number of connections, distance of each leg, etc.)


Thoughts?

ekaneti

I am a strong advocate of doing one stop thru pax first before connections

Branmuffin

Quote from: ekaneti on November 12, 2009, 12:24:38 AM
I am a strong advocate of doing one stop thru pax first before connections

Just to clarify, what do you mean by that?

ekaneti

Quote from: Branmuffin on November 12, 2009, 12:35:08 AM
Just to clarify, what do you mean by that?



Right now we can schedule a plane from ORD-DEN-LAX, but not carry pax on ORD-LAX via DEN. We should try that first before we have connections.

Connecting pax are one of those "becareful what you ask for" things

Tivec

I thought you already could carry pax ORD-LAX via DEN by making the stop at DEN a technical / refueling stop only, or am I misunderstanding you?

ali5541

Quote from: Tivec on November 12, 2009, 01:01:48 PM
I thought you already could carry pax ORD-LAX via DEN by making the stop at DEN a technical / refueling stop only, or am I misunderstanding you?

you can, and it will remain in the game!
Member since July-2007

TommyC81

Quote from: ekaneti on November 12, 2009, 12:24:38 AM
I am a strong advocate of doing one stop thru pax first before connections

Well, that would be more or less what I'm saying apart from the part where I suggest ways to limit the calculations needed to process ALL connections. Having truly connecting flights for all possibilities would likely require more processing power than all users can supply during a 30 min gameday.