Page 1 of 1

any optimal limit of bots served area or amount?

Posted: Sat Jan 26, 2019 3:12 pm
by Dixi
When I have 10-20-30 roboports and hundreds, maybe thousand of bots flying around them, serving some dedicated not big factory, like red engines production, or yellow bottles, every thing works perfectly.

But when I build some big structure covered by one drone area, like big ore to steel plant (11k steel/min), or full cycle green chip factory near iron/copper deposit (described later), that area need a lot of drones (4000-6000) and 100+ roboports.

1st time I noticed described above situation in following design.
I have a green chip factory that is located between rather big copper and iron deposits.
8 speed-3 beacons per factory, productivity-3 inside factories.
It produce about 53k green chips per minute.
206 roboports, 6300 flying bots (usually about 4500-5500 bots busy).
Adding a few more green chips factories usually leads to zero free bots, and could be solved only by adding some more roboports. Adding a few more hundreds of bots does not help much.

Whole base is build as separate specialized factories served by dedicated roboports areas, and supplied by trains.

I have feeling that it uses too much bots/roboports related to it's size, compared to other separate production areas where about 2k bot's and less amount of roboports usually enough.

I wonder, it it too big in whole, so bots fly too far on average trip?
If I split it to 3 separate areas:
1. green chips
2. copper mining & smelting
3. iron mining & smelting
with requester-stack inserter-passive provider on borders, will it be more effective? Or no?

In other words: is it effective enough to build big drone area, feed it with tens of thousands of drones, or it's more optimal to split that big area to several smaller drone areas?

Re: any optimal limit of bots served area or amount?

Posted: Wed Jan 30, 2019 1:43 am
by Frightning
afaik, modularizing the logistics robots is more efficient, not only for practical reasons, but for game performance reasons too. I seem to recall the devs stating that the algorithm for assigning jobs to logistics robots grows roughly n^3 relative to number of active bots within a given logistics area.