News:

SimuTranslator
Make Simutrans speak your language.

Factory production chains and urban delivery locations

Started by Sarlock, November 18, 2012, 06:10:29 PM

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

Sarlock

To feed off of the discussion in the other forum regarding factory production and urban consumers, I'd like to raise the thought of beginning to rebalance the pak128 factory production chain to require more urban factories to consume end products.  Each chain should be looked at individually to see how it should be handled, some might be best kept as they are.  As one of my projects I'd like to create more 1-2 tile urban factories/retailers (like the SFC) to consume all of this production.  A factory should have to deliver to multiple urban retailers in order to have its production fully consumed.

This has the following effects:

1) Creates a more realistic production chain
2) Challenges the player to plan out more complex urban goods delivery networks
3) Doesn't litter the countryside with as many primary raw material producers (though maybe some players like this)
Current projects: Pak128 Trees, blender graphics

grampybear

This is a wonderful idea. I have played this way for a long time, changing some of the wonderful eye-candy com buildings we have in pak into end user factories,etc; fast food chain, restraunts,office supply, small convient stores. The early hurdle I ran into and still do is the small amounts these places require. As talked about in an other thread the way the game loads vehicles, in this case trucks- even small panel ones, is that the vehicle looks for one item to one destination, most of the trucks carry more than a small business requires. Loading to multiple stops becomes a real challange. To some extent I have worked around it by using warehouse hubs, like a wholesale food distributor where large amounts of goods come in and then are sent out to the smaller stores. This works only fair. Distribution is your hurdle here.

VS

Yes, this would be ideal.

However, I fear the game does not really support this kind of chain linking, only accidentally. Feel free to correct me with solid evidence, though ;)

My projects... Tools for messing with Simutrans graphics. Graphic archive - templates and some other stuff for painters. Development logs for most recent information on what is going on. And of course pak128!

IgorEliezer

Quote from: Sarlock on November 18, 2012, 06:10:29 PM
2) Challenges the player to plan out more complex urban goods delivery networks
Perfect! This would "force" the player to use non-conventional solutions for urban traffic, e.g. elevated bridges, multilane avenues and underground roads.

Sarlock

I am not too familiar with the inner workings of how factory chains are created and expanded so I will have to do some experimentation with the process to gain a deeper understanding of the best way to implement this idea.

When a factory chain is initially created, it does attempt to build the entire chain from raw materials to end consumer and it does create a bit of balance between the different parts of the chain based on production rates.  If the chain was inverted to have a production level balance of 2 raw material producers feeds 1 factory feeds 4 urban retailers, would the initial factory setup approximate this?  Would the urban factories(consumers) likely all be created in the same city?  If there were 4 different types of end consumer for that same product, would the game randomly select from the 4 different end consumers or use the same end consumer each time?

Example:
2 cattle farms produce meat
1 meat packing plant produces canned meat
4 supermarkets consume canned meat

If we had 4 different types of supermarkets, would the game choose randomly from those 4 or use the same supermarket each time for this chain?
Current projects: Pak128 Trees, blender graphics

VS

Chains are always built top-down. What is the probability of linking to another factory, remains the question on which this scheme hinges.

My projects... Tools for messing with Simutrans graphics. Graphic archive - templates and some other stuff for painters. Development logs for most recent information on what is going on. And of course pak128!

Sarlock

I'll run some controlled experiments and see what kind of data I can gather.
Current projects: Pak128 Trees, blender graphics

ӔO

from my experience, the final consumer always spawns in inadequate numbers.
My Sketchup open project sources
various projects rolled up: http://dl.dropbox.com/u/17111233/Roll_up.rar

Colour safe chart:

sdog

you ought to contact AvG, he did some experiments with exactly this in pak128.britain. Set consumption to very small values and created large low density consumer networks. He tends to do things very systematically and might have some data or at least some experience to share.

Dwachs

From looking at the code, Imho the program should
(1) only create a full chain to produce one particular good for a consumer
(2) tries to connect to existing factories with surplus production before spawning new factories

That said, if you have end-consumers with low consumption rates, then you will get more consumers than producers on the map.

The other way round, consumers with high consumption rate, will spawn a lot of new industries on creation.
Parsley, sage, rosemary, and maggikraut.