Factorio Forums

Any flaws in this design? I am here at 126/138= 91,3 %

14 miners, 10 chests, 2 power poles, no unused space

Looks good. I suspect that is the optimal solution for medium power poles. There *might* be a crazy tessellation using substations that is 1-2% better, but I doubt it.

I've a program on iPad, called iOrnament. Besides that it is a really good program, this program has a fantastic documentation about tessellation, ornaments and tile and explains the mathematics behind them in a very easy way.

Here some of contents (translated into English as good I can):

There are 4 types of elementary symmetries:
- movement
- rotation
- mirroring
- slide-mirroring (?) - this is a mirroring with a following movement in the direction of axis of reflection

There are 17 ornament groups, which follow from the possible combinations of symmetries. They have names, like P1, Pm, C4 (used in crystallography) and a orbifold symbol (which is used by mathematicans) like 3*3, 0, ** or *x.

Then for quadratic tiles (which factorio uses in general) we have some special rules: every tile is a copy or mirror of the original. Then we have 14 possible ways of combining them, or 31, if we also allow the rotation of the original tile.

There are 3 mirror groups, which you can see in a kaleidoscope and which also can be seen in the above examples. The sum of the angles must be 180 degrees: 60,60,60 (P3m1), 90,45,45 (P4m) and 90,60,30 (P6m).


There are more operations like stretching, but that's not important for factorio.

Why I'm writing that? Well, if we could implement the mathematics behind that into placing repeated numbers of the blueprints, then we can create things like above automatically.

What's needed?
- definition of a blueprint
- original placement of first blueprint and how it is rotated or mirrored.
- one of the 17 (14?) orbifold symbols to define the type or repetition
- a statement/function for stoping the process (is there water, forest?)
- a statement/function for continuing (is there copper? Is the previous built up complete?)
- a mechanism to request the needed items for that
- a "flying" roboport (something, which moves with the growing construction to allow the buildup) and
- some physical instance, which creates ghost buildings from the blueprints by the given rules.
- optional: rules/functions what has to be done before placing (e. g. Deconstruct all) and after (cut power connection on that pole). And rules, which allow to change the blueprint ... or the whole program.


I think some (not all) of the above mentioned is essentially important to create self replicating structures. And what wonderfully would it look to generate big structures just from a simple blueprint and some rules?

Most vector image editors, like Inkscape and Illustrator, also have tools for all of the different symmetry groups, and more.

Also, don't assume that a repeating pattern is the optimal solution here. There is such a thing as tiling that can never repeat, such as penrose tiles.

I did the hard work for you



This is 3 'wasted' space per 3 miners, or 90% efficiency.

I was thinking of a method to use substations and four miners facing each other, but it's hard to get a good setup.

sparr wrote:To maximize mining speed you want to maximize miner density. Zig-zag belts waste space (and are slower!), and so do power lines.

In your example, for every 36 squares of miner there are 7 squares of conveyor and 1 square of power, so you've covered 82% of the ground with miners.

Here's a layout that covers 86%.

Cheers!

vanatteveldt wrote:This is 3 'wasted' space per 3 miners, or 90% efficiency.

To be fair, we should probably be including the footprint of a roboport in the logistic chest solutions here, because in a large enough deposit the roboport will have to sit on some of the ore. The roboport uses up 9 spaces out of 2500, so I think this is 89.676% efficient if tiled indefinitely.

The setup with 3 electric drills gives 0.5 items/sec for each drill, so that's 1.5 items/sec on 100 tiles.

You can also place 4 burner drills around 1 chest, this is tileable.
Burner drills give 0.35 items/sec each.
This setup uses 17 tiles, so 1.4 items/sec/17 files, 8.235 items/sec/100 tiles.

With 3 speed module 3 modules in each electric drill you get 150% speed bonus, but that's still only 3.75 items/sec/100 tiles.
The downside of this is the manuel filling the drills with fuel.

Another option would maybe be to use beacons with the electric drills. Who wants to try?

Electric drills also mine one row around them, so you get a 5x5 mining area with a 3x3 footprint (+ infrastructure). Additionally those areas can overlap.

sparr wrote:

vanatteveldt wrote:This is 3 'wasted' space per 3 miners, or 90% efficiency.

The roboport uses up 9 spaces out of 2500, so I think this is 89.676% efficient if tiled indefinitely.

Roboport uses 16 spaces. Not many deposists are that big though even with mods that increase their size, low frequency and big patches from settings.
And a 4x4 entity doesn't tile that well with 3x3 entities so you might get even lower space efficieny if you have to remove 36 tiles of miner to squeze in a roboport. You might have to redesign the tiling pattern to not lose 4 miners when a roboport is inserted.

There is no need in roboport, belts can be used instead of Chests.

Altren wrote:There is no need in roboport, belts can be used instead of Chests.

Nice! 88.9% space efficiect (48/54).
But it puts more items on one side of the belt than the other so you can't get full belt capacity because I think all upwards facing miners put on the same side (same for downwards facing) on vertical belts. And on each belt you are either only placing miners facing either up or down. Would be even better if you could alternate between miners facing up/down to fill both sides of the belt equally.

Qon wrote:

Altren wrote:There is no need in roboport, belts can be used instead of Chests.

Nice! 88.9% space efficiect (48/54).
But it puts more items on one side of the belt than the other so you can't get full belt capacity because I think all upwards facing miners put on the same side (same for downwards facing) on vertical belts. And on each belt you are either only placing miners facing either up or down. Would be even better if you could alternate between miners facing up/down to fill both sides of the belt equally.

It is 90% efficient, same as in previous solution (3 tiles wasted for each three mines, 27 out of 30 tiles are mines or 54/60 if count both sides).

Also I made balanced up/down-mines solution, but mines always place ore on the left side, so it seems, that there is no way to make it provide balanced output.
Same problem when trying to use horizontal layout - both left and right-sided mines place ore on lower line.

Altren wrote: It is 90% efficient, same as in previous solution (3 tiles wasted for each three mines, 27 out of 30 tiles are mines or 54/60 if count both sides).

Also I made balanced up/down-mines solution, but mines always place ore on the left side, so it seems, that there is no way to make it provide balanced output.
Same problem when trying to use horizontal layout - both left and right-sided mines place ore on lower line.

Ooops, yeah failed my calculations there. Thanks for correction.

;/
Well you could make something like a ~100 tiles high chain of miners if 2 of them output on the same lane always before you decrease efficiency. If you go from center outwards you could double that.

(math probably wrong lol, I'm tired and doing it in my head: blue belt 40 items/s. 1 lane = 20 items/s. 1 miner outputs 0.52 ores/s or something, so almost 40 miners in a row before one lane is filled. miner side length 3*40 = 120)

I didn't realize a miner could push items onto an underground belt from the "back". Good to know!

Really cool.

Grr, should have saved a draft...

Anyway! Lost the post I was going to do because the site logged me out, so here's the short version:

The alternating design in Altren's post is bad if you're using belts. It ends up taking 4560 tiles to fully output a blue belt, but the simple balanced miner design with underground belts only takes 2394 tiles to output a blue belt.

However, if you're using bots then the alternating design is superior. How do you get a robotport in it? This way:


There's a great 6x4 area in the middle so you only lose out on 2 and 2/3 miners.

Shown on an Angelbob's map because I'm lazy in a weird way and didn't want to change my mods but happily spent more than an hour on fiddling with the design.

Altren wrote:There is no need in roboport, belts can be used instead of Chests.

CompactMine.jpg

I like

Created a blueprint for it

Zourin wrote:sadly, I still prescribe to the 'low, slow' and steady' of lining them up by their mining radius, rather than their physical space. More longevity, same amount of resources, less power drain, less pollution.

But still.. an interesting discussion nonetheless.

Depends on what you want to do I suppose - if you want to stripmine a small field to clear space or pack up and move on, a dense cluster with speed modules makes sense. If you've got a large field and want to make it last, going wide and adding production modules maximizes the time the mine will be active. I prefer the somewhat slower approach.