Whats a Good Number of Buses?

[Zutty]

Hi all,

I've really been dragging my heels lately. I'm quite busy with work, and I've only just finished Mass Effect (which was awesome by the way!)... however I'm hoping my AI development can get back on track now.

I keep saying that I'm close to release, but despite the fact that my AI is quite functional it is in desperate need of balancing. To be honest I'm dissuaded by the impressive numbers that other people are getting!

I think my biggest problem right now is over-production of buses. After 10 years or so my AI can be running anything from 300-500 buses, which is clearly WAY too many. Running costs alone take a huge chunk out of the profits, but the killer is the cost of actually buying that many vehicles and later in the game the cost of replacement vehicles for auto-renew (I daren't even think about the cost of upgrades when new models become available).

I could just reduce the weights in my fleet size equation, but I think the underlying problem it that its spreading itself too thinly, and reducing numbers of buses across the board will just exacerbate the problem. I'd rather have a sound policy for estimating the number of buses that a town is actually capable of supporting.

At the moment I'm using the sum of the cargo acceptance of every bus stop in a town, but this gives too little for large towns and too much for small towns. I tried a square rule but this goes too far the other way. Any suggestions?

Thanks.

[Ralph]

My method is to guess a low figure when building the route (2 + distance/100 + sum of acceptance/100) and then check the stations on a regular basis and add more buses if there are lots of passengers waiting, at the same time check for buses not making a profit and sell them. It's not perfect yet, but works well enough. This also means it will adapt as the town grows.

[Yexo]

I'm doing something like what Ralph does. Guess a good number to start with (always 6 , and later on check if there are passengers waiting, if yes, build some new busses (depending on amount of waiting passengers). I also have a limit of busses per station, or the busses will flood my stations.

[Finaldeath]

I'd put a hard limit on single stations (higher perhaps on bus stops), since the simple mechanic of going in and out of a bus stop means anything over 4 or so waiting vehicles really just halts the efficiency of the stop, so the number of passengers never goes down fast enough.

Depends on the distance of course, but those big groups means the buses are not dropping off their cargo fast (making no money) and not picking up passengers fast (gets a worse rating), all in all, bad

I need to figure out the optimum, but am looking at water AI still ATM.

[yorg]

I'd put a hard limit on single stations (higher perhaps on bus stops), since the simple mechanic of going in and out of a bus stop means anything over 4 or so waiting vehicles really just halts the efficiency of the stop, so the number of passengers never goes down fast enough.

ya, i'd say this is the number 1 problem with robotAI at the moment, if a station services an industry with a very high production no number of road vehicles will be able to maintain a half-decent station rating, which means the AI keeps building vehicles that spend most of their time queuing for the station. I think I am just going to add some functionality to add an extra station building for every, say, 20 vehicles.

with regards to the original topic, robotAI only concerns itself with a station rating target - it builds one vehicle to start, and if the rating ever drops below 55% (for the moment, at least) it adds another one. It actually works surprisingly well, at least for industries. except in the case of industries with very high production, it seems to all go bananas once it's reaching 300 units a month.

[athanasios]

number of vehicles and distance or more accurately time to go and return should be taken into consideration.