|
Nalesh posted:I've forgotten how painful it is to not have robots, any suggestions on an early bots mod? Industrial Revolution has you unlock steam powered construction bots real early.
|
#
?
Jun 23, 2022 09:22
|
|
|
|
| # ? Apr 24, 2024 02:29 |
|
|
In the same vein, the one I always use when doing a modded playthrough is FasterStart It gives you modular armor with 25 slow construction bots, nightvision and a single exoskeleton. It let's me use bots to help me with the annoying early game building by letting my cut, copy and paste. They are slow as hell though which is why I like these instead of other mods that give you some crazy fast bots right from the start. I will want to force myself to actually go for whatever type of bots are available in the game I'm running. Speaking of which, I'm doing a Bob&Angels playthrough and I unlocked the crawler bots. These have unlocked a lot earlier than I thought and they seem perfect for what I want. Trying to setup any type of mall for these major overhaul mods is always a massive pain due to all the different types of components you need. Is it recommended to use these things? I played around with them and they are even slower than my faster start bots, but speed really isn't much of an issue for the spot where I'm using them. I have unlocked the normal bots too but that produciton chain is still miles off and I'd like to automate a bit more than just my belts and inserters.
|
|
#
?
Jun 23, 2022 10:45
|
|
|
I mainly meant the logistics side of bots, but I found one called robot world and it seems to be what I want. I tried Kizrak's robots but god drat they are not worth it, insanely slow and only having 3 spots on requester chests.
|
|
#
?
Jun 23, 2022 11:57
|
|
|
KirbyKhan posted:Anyone try out that Spaceblock thingamajig? Seems like instead of generating things from the sea it's more about generating positive feedback loops and skimming the excess for the next, higher tier, feedback loop. I tried Spaceblock, and apart from a few recipes to duplicate ores it seemed to be an entirely vanilla game progress. Maybe that is your thing but I'm getting bored with vanilla Factorio.
|
|
#
?
Jun 23, 2022 12:07
|
|
|
 I love this design so much, I just wish bots worked in the factorissimo building without having to take up space, it breaks the efficiency
|
|
#
?
Jun 23, 2022 22:48
|
|
|
I have made a brown algae in seablock. Hrmm, I should stop here. I might dig up my old chemistry book to look at my dimensional analysis formula notes. Oh no
|
|
#
?
Jun 23, 2022 23:51
|
|
|
Tempted to do double sided trains for my LTN network this time, how do ya'lls stops look?
|
|
#
?
Jun 24, 2022 00:38
|
|
|
I’m always tempted by LTN, but my experience has consistently been that it fixes all the regular problems with my network, then replaces them with exciting new problems. On the other hand, it’s certainly very cool.
|
|
#
?
Jun 24, 2022 01:09
|
|
|
Nalesh posted:Tempted to do double sided trains for my LTN network this time, how do ya'lls stops look? There is really no need for that, that'll just increase congestion and complexity for no gain. All you need is a simple loop to obsolete double sided trains.
|
|
#
?
Jun 24, 2022 04:07
|
|
|
Nalesh posted:I mainly meant the logistics side of bots, but I found one called robot world and it seems to be what I want. I tried Kizrak's robots but god drat they are not worth it, insanely slow and only having 3 spots on requester chests. Try logistics drones. Makes road-based earlyish game logistics.
|
|
#
?
Jun 24, 2022 04:36
|
|
|
Here's an interesting "belt set" question for the algorithms people in the thread. For various reasons*, I was interested in "settling" a belt, so that its contents are present as far to one side as possible (let's say the left, without loss of generality). For example, let's say we have 8 belts. If, say, the first three belts are full, the fourth is partially full, and the rest are empty, the belts are settled: if the first two belts are full, the third is empty, and the fourth is partially full, the belts are unsettled. There is a simple way to settle a set of belts: Repeatedly insert "zippers" of splitters with output priorities to sort the set of belts one at a time. A "zipper" consists of a set of splitters with output priorities towards the left of the belt set, starting at the end. For example, if we have 4 belts, first there would be a splitter between 3 and 4 preferring 3 as output. Then, one between 2 and 3, and finally one between 1 and 2. This would ensure that either 1 is full or every other belt is empty. You can then apply such a zipper to the remaining belts except 1, and so on. In the end, you will have used (n - 1) + (n - 2) + ... + 1 splitters, which comes out to (n-1)(n-2)/2. The question that has been bothering me is: Is there a better way to do this? In other words, can a set of belts be settled while losing only trivial amounts of throughput and using fewer splitters than this? I thought about it a little, but I still have no idea. I tried a divide-and-conquer approach, but a good approach to the merge still eludes me. * The reason is a little embarrassing: I have been creating train stops to do assembly, but am short on prod 3 modules. So, I wanted to divert most goods to the first set of assemblers I built, when I actually had a full inventory of prod 3 modules, and then the rest will go to later assemblers which are probably missing modules. Don't judge me.
|
|
#
?
Jun 26, 2022 02:48
|
|
|
I routinely use output priority on balancers to ensure certain things get resources preferentially. I think that’s what you’re asking, anyway. The real answers though, is to increase production of the limiting resource. 
|
|
#
?
Jun 26, 2022 02:53
|
|
|
Oh yeah, practically it doesn't really matter, since if all the belts are full it doesn't matter how they're ordered. But I think it's still an interesting problem.
|
|
#
?
Jun 26, 2022 02:57
|
|
|
Sure, a few ideas: 1. Use logistics robots to bring resources to the assemblers instead. 2. Stop overthinking it and wait for a few more prod 3s to be made. 3. Only put one zipper per belt. It'll use fewer splitters at the cost of slightly less perfect balancing.
|
|
#
?
Jun 26, 2022 02:58
|
|
|
nrook posted:The question that has been bothering me is: Is there a better way to do this? In other words, can a set of belts be settled while losing only trivial amounts of throughput and using fewer splitters than this? I thought about it a little, but I still have no idea. I tried a divide-and-conquer approach, but a good approach to the merge still eludes me. You can reduce the splitter count if you allow swapping of belts (I worked out 4 belts with 5 splitters) but then you're adding underground belts + extra space to swap the belts. I haven't worked out the general solution for an arbitrary power of 2 yet to see how it scales with splitters, but I suspect there's still a quadratic/(N choose 2) scaling for total swap + splitter operations
|
|
#
?
Jun 26, 2022 05:14
|
|
|
I think your formula is wrong - your succession of smaller zippers actually requires n(n-1)/2 splitters in total for n belts. (Most easily seen with 3 belts - your formula would suggest that this is possible with a single splitter, which it clearly isn't.) I think that's the best you can do in the general case, where inputs could be completely unbalanced and you might need to move the k rightmost lanes all the way to the left hand side. But for the special case where inputs are approximately balanced (due to coming from a train stop unloading), you can almost certainly do better just by leaving off most of the later zippers - the earlier zippers will end up doing that work.
|
|
#
?
Jun 26, 2022 05:23
|
|
|
Jabor posted:I think your formula is wrong - your succession of smaller zippers actually requires n(n-1)/2 splitters in total for n belts. (Most easily seen with 3 belts - your formula would suggest that this is possible with a single splitter, which it clearly isn't.) Oh yeah you’re right wtf. I looked up the sum of 1 + 2 + … + n and saw n(n-1)/2. But I didn’t see that this is obviously wrong— it’s n(n+1)/2. I am pretty sure I was half asleep and just turned the formula in my head into the one for getting the number of distinct pairs in a set of n entities. And yes, there are definitely cheaper “good enough” solutions. Another one is that if you just removed part of the stuff on the first belt, you only need one “reverse zipper” (starting at the beginning this time) to resettle the belt, I think. man in the eyeball hat posted:You can reduce the splitter count if you allow swapping of belts (I worked out 4 belts with 5 splitters) but then you're adding underground belts + extra space to swap the belts. I haven't worked out the general solution for an arbitrary power of 2 yet to see how it scales with splitters, but I suspect there's still a quadratic/(N choose 2) scaling for total swap + splitter operations I’m curious to see this solution! I had planned on calling swaps free (even though in actual Factorio they do cost underground belts and resources) just because otherwise the problem seemed impossible to analyze.
|
|
#
?
Jun 26, 2022 11:18
|
|
|
Hi there, I was hoping I could get some advice here. I've put about 80 hours into Factorio on and off over the last couple of years but I have a problem where when the infrastructure reaches a certain level of complexity it kind of gets beyond me. This time round I've been trying to understand the systems better so I can play better. Trains have been a massive Achilles heel for me, I find it really hard to grasp how the signalling system is meant to work so I tend to accordingly make them as simple as possible, they usually end up loops that don't intersect and have one train on them. This time though I've been trying to understand it better so I can actually use trains to their full effect, so I've been experimenting with more complex setups. So I put together this cloverleaf leaf looking design with four trains, all on the same track, that intersects and has four stations. I put the signal lights at the beginning of each intersection and it didn't work at all, one of the trains bumped into another and the others all got into a Mexican standoff with each other where none were willing to move:  Next I put some more signal lights to cordon off each of the stops and suddenly everything runs in perfect harmony, no crashes or significant downtime with trains stuck, everything moves at a reasonable click to get to the assigned stations:  I'm glad it kind of works, but I don't really understand why it does? Can somebody give me an explanation so I know how to make sure the trains run on time in future?
|
|
#
?
Jun 26, 2022 22:01
|
|
|
In the first screenshot, the bottom train can't go forward because the left train is occupying the purple block. The center train can't go forward because the left train is occupying the purple block. The left train can't go forward because the center train is occupying the light blue block in front of it. A train will wait at a signal until it can reserve and enter the block in front of it. A train will wait at a chain signal until it can reserve and enter all the blocks it must enter to pass a regular signal. Chain signals look ahead and "chain" reservations, regular signals the reservation stops with that block. This sounds much more complex than it is, I recommend you play with it a bit and you'll probably understand it fairly easily. What you did to "fix" it doesn't actually properly fix the system. If you added more trains it would eventually grind to a halt again, before it should if properly signaled because you have a huuuge single purple block covering the whole intersection and any train occupying any of it is blocking other train from entering any of it. Chain signal before every entrance to an intersection, regular signal after every exit. That's how you make them actually consistently work. That way, trains will not enter an intersection until they are able to also exit it, and thus they won't sit in the middle of it jamming everything up. In your first example, even though it doesn't look like it, the left train is effectively sitting in the middle of the intersection. That plus leave enough space between intersections and putting enough regular signals along those open areas will keep your rails running smoothly, generally speaking. It certainly will let you get enough understanding to realize why things aren't working or efficient in the potential more complex/dense scenarios. Generally speaking, the best beginner rule of thumb is the only make 3 way intersections, not 4 ways. K8.0 fucked around with this message at 22:32 on Jun 26, 2022 |
|
#
?
Jun 26, 2022 22:25
|
|
|
Alright thank you, so I'm dicking around with it a bit more and added the Rail chain signal at the entrances and it seems to have made everything much faster.
|
|
#
?
Jun 26, 2022 22:53
|
|
|
The next thing you would want to do would be make sure that there is a regular signal 1 train length away from each entrance and 1 train length away from each exit. That way, if 2 trains are close behind each other, they won't have to wait as long. Right now if you had another train directly behind the one in the center, it would have to stop until that train made it all the way out of the purple block, at the top left. By putting a signal 2 car lengths past the one at the exit, it would only have to wait for the first train to clear that first block before it could also proceed. You can set the length of the ghosts show in the options to match the trains you're currently using/planning.
|
|
#
?
Jun 26, 2022 23:07
|
|
|
If I'm finding vanilla progression to be a little too barebones for complexity-obsessed rear end, would K2 be my best option? I've given B+A and SE a look and was immediately put off by the needlessly complex recipes for basic automation items. I do wish I could keep the steam powered automation stuff from Bob's though, that part at least is really neat.
|
|
#
?
Jun 26, 2022 23:36
|
|
|
Bloody Pom posted:If I'm finding vanilla progression to be a little too barebones for complexity-obsessed rear end, would K2 be my best option? I've given B+A and SE a look and was immediately put off by the needlessly complex recipes for basic automation items. Yes, K2 is probably the best overhaul mod if you want 'vanilla, but more of everything'. If you want steam powered machines, Industrial Revolution is supposed to be good, but I refuse to touch that because of its dev.
|
|
#
?
Jun 26, 2022 23:54
|
|
|
Also if I'm reading this right, you can build Factorissimo buildings inside themselves? Forget spaghetti hell, we're doing non-euclidean matryoshka hell. Bloody Pom fucked around with this message at 00:09 on Jun 27, 2022 |
|
#
?
Jun 27, 2022 00:04
|
|
|
khwarezm posted:Alright thank you, so I'm dicking around with it a bit more and added the Rail chain signal at the entrances and it seems to have made everything much faster. K8.0 posted:Chain signal before every entrance to an intersection, regular signal after every exit. Let's do this. And then some more. Start with putting a regular signal on every exit. Technically your test track may only have one direction but I'm going to reverse it to make it look more right hand drive. (Otherwise you have a train on the left going upwards and downward on the right or the top one is going to the right and the bottom one to the left  )I start with regular signals on every exit.  Now some chain signals on every entrance.  Then we can split the crossing into two parts, allowing one train to cross from the left and one from the right at the same time.  One more split.  Beautiful. If you have two trains on the parallel tracks, they may use the crossing at the same time. Again: K8.0 posted:Chain signal before every entrance to an intersection, regular signal after every exit.  Notice the blocks again. Every train - no matter the direction it wants to take - won't block any other train with their block-reservation if they don't cross paths.
|
|
#
?
Jun 27, 2022 00:20
|
|
|
Tamba posted:Yes, K2 is probably the best overhaul mod if you want 'vanilla, but more of everything'. Giving it a look and one thing I'm especially excited to muck around with are the loaders. Being able to load/unload belts at full speed should make rail depots less of a bottleneck compared to even fast inserters.
|
|
#
?
Jun 27, 2022 00:28
|
|
|
khwarezm posted:Hi there, I was hoping I could get some advice here. Basically, you use signals to split the track into separate chunks called "blocks", and then each block can only contain one train at a time. This prevents trains from running into each other or colliding, because a train won't enter a block that another train is already in - the train essentially reserves a block for its exclusive use, and no other train can enter that block from any direction until the first train has left. You're treating signals as stoplights, where you just put them at a place where you want a train to stop and wait for the route to clear. But because of this absolute reservation system, you also have to place a second signal to mark the size of the area that you want to be clear. And if it's at an intersection, you'll have to place signals at the other rails coming in and out of the intersection too. You have to place signals not just at the entrances, but at the exits as well. Think of it as drawing a box with signals, and once a train enters that box from any direction, no other train will be able to enter until that first train has passed one of the signals marking the exits of the box. The colors that show up on the rails mark the sizes of these blocks. For example, in your original very first screenshot, all four cloverleaves were in the same block - all of the purple rail was contiguous with no signals splitting it up, which meant that only one train could be on any of the purple at a time. This is why things are better in your most recent screenshot: since you have signals at every entrance and exit of that square of track at the very center (as you can tell from the new color), you've created a dedicated block for all the rail crossings. The rest of your cloverleaf is all straight one-way rail with no crossings, and it's difficult for things to get too hung up on that. Before you redid the intersection signals, the cloverleaves weren't properly split into different blocks; now they are. There are other important effects that come from this absolute reservation system. For example, the distance between signals affects how closely trains can follow each other, because only one train can be between a set of signals at a time - so if you've got a really long distance between two signals, then once one train enters that stretch, no other train will enter until the first one has passed the exit signal. And if a second train is waiting to enter, then it's occupying a block itself, and no other train will enter that block until it moves on. If your blocks are big because the gaps between signals are long, then the trains lining up and waiting will take way more space than the size of the trains themselves. And if you're only placing signals at important places, then they'll all be sitting and waiting in important places. So even on a long straight stretch of rail with no intersections, you'll want to place signals every so often to help smooth out the movement of your rail network That's been another contributor to your issues - since you don't have any signals between your stops, only one train can be between your stops at a time. As it's currently set up, each cloverleaf can hold up to four trains: one at each stop, one between the two stops, and one in the stretch between the entrance of the leaf and the entrance of the first stop. Even though there's so much space between the entrance of the leaf and the entry to the first stop, only one train can be waiting in that space at a time - no other train will be able to enter the cloverleaf until that stretch of rail is clear. Lastly, there's chain signals. They exist to solve a fairly basic problem with basic block signaling: you have to place signals at the exits of an intersection, but if the next block after the intersection is occupied, those signals at the exits can become red lights and can cause a train heading that direction to stop inside the intersection to wait for the next block to clear. Chain signals are a very simple solution - they work just like normal signals, except that in addition to checking the status of their own block, they look ahead and check the status of the next signal. If the next signal is red, then the chain signal will turn red too, even if its own block is empty. So if you place a chain signal at the entrance to an intersection, then it won't let a train enter the intersection unless the exit to the intersection is also clear. And in cases where there's multiple exits, the chain signal will actually check which exit an approaching train wants to use, and will check only that exit when deciding whether or not to let the train through.
|
|
#
?
Jun 27, 2022 00:30
|
|
|
RabbitWizard posted:You see the colored lines, those are the "blocks" that get reserved. As soon as one train reserves a block, no other may enter. See the problem with your crossing? It's all one block. Thanks, that's a big help, especially since you went through the trouble of screenshotting all this. I'll probably be back at some other point to ask about robots and circuits.
|
|
#
?
Jun 27, 2022 00:47
|
|
|
Are there any mods that add a (presumably late-game) more C&C-like resource gathering system? i.e. You drop a dock on/near an ore patch and an automated harvester vehicle goes around hoovering up ore and depositing it back at the building, which then feeds it into your logistics network? I ask because playing Tetris with electric miners gets pretty tedious eventually.
|
|
#
?
Jun 27, 2022 01:00
|
|
|
Bloody Pom posted:Are there any mods that add a (presumably late-game) more C&C-like resource gathering system? i.e. You drop a dock on/near an ore patch and an automated harvester vehicle goes around hoovering up ore and depositing it back at the building, which then feeds it into your logistics network? I ask because playing Tetris with electric miners gets pretty tedious eventually. It's been a while so the technical aspects may not be totally correct (any more), but last time I player it was kinda like in C&C. To automate it, you put down a dock,
|
|
#
?
Jun 27, 2022 02:08
|
|
|
RabbitWizard posted:You "want" https://mods.factorio.com/mods/Earendel/aai-programmable-vehicles After my experience with SE forcing AAI down my throat I kinda want to stay away from it. I did find another one that adds mining drones but I'm unsure if it'll play nice with KR2. Does anyone have a good setup for siphoning output from a smelting line into a bus, without leaving an ugly gap in the furnace/inserter chain? I'm sure I can figure something out with splitters/underground belts but I don't think the undergrounds have enough reach until red belts. Also, how do y'all like to handle glass production in KR2? I've seen some blueprints use filter splitters to pull stone off the end of a brick line, but I think I'd rather avoid clutter by giving glass its own line and feeding stone directly from the bus into the crushers. Do crushers even serve a purpose beyond stone>sand and garbage disposal? Bloody Pom fucked around with this message at 03:42 on Jun 27, 2022 |
|
#
?
Jun 27, 2022 03:34
|
|
|
Bloody Pom posted:Do crushers even serve a purpose beyond stone>sand and garbage disposal? Nope. Other stuff will use sand though, so you'll need crushers in a few places.
|
|
#
?
Jun 27, 2022 03:52
|
|
|
Taffer posted:Nope. Other stuff will use sand though, so you'll need crushers in a few places. Good to know. Might have to try setting up crushers near my stone mines and bussing sand rather than producing it onsite.
|
|
#
?
Jun 27, 2022 04:00
|
|
|
I wouldn't recommend that. The recipe produces multiple sand per stone so it will be far less dense on a belt. The machine is pretty large but it will simplify transport a lot to have it produced on site.
|
|
#
?
Jun 27, 2022 04:04
|
|
|
nrook posted:I’m curious to see this solution! I had planned on calling swaps free (even though in actual Factorio they do cost underground belts and resources) just because otherwise the problem seemed impossible to analyze. I haven't built it in game, but I believe this should work/was backed up by some simulations I ran outside the game: You have 4 belts, numbered from left to right as 1 to 4. Settling is a priority splitter with priority to the leftmost belt. Swapping is reordering the belts with some extra space and underground belt shenanigans. Settle belts 1,2 Settle belts 3,4 Swap belts 2,3 (now 3', 2' respectively) Settle belts 1,2' Settle belts 3',4 Settle belts 2',3' I'm actually curious to figure the general answer out now. The naive approach feels similar to a bubble sort, but it's not really a sorting problem because you can merge entries in the list.
|
|
#
?
Jun 27, 2022 04:13
|
|
|
Bloody Pom posted:After my experience with SE forcing AAI down my throat I kinda want to stay away from it. I did find another one that adds mining drones but I'm unsure if it'll play nice with KR2. Mining Drones and K2 work together without any issues, but it's more 'Age of Empires' rather than C&C Tamba fucked around with this message at 06:38 on Jun 27, 2022 |
|
#
?
Jun 27, 2022 06:36
|
|
|
Bloody Pom posted:Giving it a look and one thing I'm especially excited to muck around with are the loaders. Being able to load/unload belts at full speed should make rail depots less of a bottleneck compared to even fast inserters. You can also get that stand-alone with the Mini Loader mod, so you can have loaders with any other overhaul mod also.
|
|
#
?
Jun 27, 2022 06:47
|
|
|
Re: Trains, I agree with the common opinion that 4-way junctions are not worth the effort for all but the slowest networks. And that you should build your network in such a way that you only use T-Junctions. Oh, and clover-leaves are a bad layout in Factorio, because you don't have bridges. If you want a conceptually simple, but slow 4 way interchange use a roundabout.
|
|
#
?
Jun 27, 2022 09:22
|
|
|
Bloody Pom posted:Also if I'm reading this right, you can build Factorissimo buildings inside themselves? There's even a start that starts you inside a factory iirc, which is real hard for obvious reasons.
|
|
#
?
Jun 27, 2022 16:43
|
|
|
|
| # ? Apr 24, 2024 02:29 |
|
|
Nalesh posted:There's even a start that starts you inside a factory iirc, which is real hard for obvious reasons. Is this in the base mod or is there an additional mod that does this?
|
|
#
?
Jun 27, 2022 16:51
|
|
