Should Construction Robots Prioritze Storage Chests ?

[jodokus31]

Construction bots take from the nearest source, because it's designed like that.

Yes, posila's post makes that clear.

If you want bots to take from the storage first, you have to ensure, that either storage is empty or storage is nearest. That results in those clunky solutions.

(emphasis mine)
Which is why the OP made this thread, and I'm in support of their idea.

But you also should consider to change the goal. Current goal is:
Storage should be used first before taking from other source and before producing more. Building can be slowed down as long you have items in storage.
Changed goal:
Storage should be used before producing more (except a certain residual). Building should be as fast as possible.

Except I'm ok with the "current goal" whereas the "changed goal" is exactly the "clunky options" I don't like? I mean, provided I don't have hundreds or thousands of what I'm building all in storage, once it's cleaned out, building then follows normal priority, so it shouldn't slow things down that long. And the "changed goal" is just a way of forcing it take from storage by letting the primary sources run out, which means it goes "slow" anyways.

I really meant, the clunky options appear, because you try to change the designed behaviour to behave different. Precisely: try to empty the storage first. But if you are fine with potential residual staying in the storage, which doesn't really hurt anyone(?), it's quite easy. limit output inserter of producer assembler to contents of logi-network. No need for circuit wires here.

Of course I see the benefit, to first empty the storage, because it sucks a bit to produce and use items, while a residual stays in storage unused. The slow down effect appears more randomly, when the storage is used at some random occasions. But I really wonder, if you are overall a bit more productive with the downside of keeping a random buffer in storage, which might happen to be used first, because its just nearer to the building location.

Koub's suggestion of reading the logistic network to disable the inserters that fill output chests with construction items when you have "enough" (rather than eg limiting the output chests) addresses that.

Yeah, if I understood Koub's post correctly, their suggestion falls in on #3 from my previous post.

Yes, it's functionally equivalent to your #3, except there's no circuit wire everywhere. You stressed the latter so infer that is a significant aspect of what makes it "clunky" to you, but it is not a necessary part. The only downside I see is the need to configure the inserters, which I think is much more fiddly and (therefore) error prone than limiting a chest. Do you see more or something else, or do you think it is just weighting (ultimately, opinion)?

Yeah, like above mentioned, i don't see the need for circuit wires. Just limit the inserter with logi-network contraints.
I really like the mod: https://mods.factorio.com/mod/AdditionalPasteSettings to set those limitiations easily.

As summary, I'm really not sure, what is preferable. I would also be ok, if storage chests have higher prio, because its nice to don't have random items in storage, which appear useless

[FuryoftheStars]

Yes, it's functionally equivalent to your #3, except there's no circuit wire everywhere. You stressed the latter so infer that is a significant aspect of what makes it "clunky" to you, but it is not a necessary part. The only downside I see is the need to configure the inserters, which I think is much more fiddly and (therefore) error prone than limiting a chest. Do you see more or something else, or do you think it is just weighting (ultimately, opinion)?

Yeah, like above mentioned, i don't see the need for circuit wires. Just limit the inserter with logi-network contraints.

Well, I had to go into the game for a moment cause I was like "what the H are they talking about don't need wires??" I'm a bit embarrassed to say I didn't know the logistic network was wireless. I've always been connecting to a robo port set to output the logistic network contents and then went from there with circuit wires. How long has the logistic network been wireless??

I really meant, the clunky options appear, because you try to change the designed behaviour to behave different. Precisely: try to empty the storage first. But if you are fine with potential residual staying in the storage, which doesn't really hurt anyone(?), it's quite easy. limit output inserter of producer assembler to contents of logi-network. No need for circuit wires here.

Maybe it might help for me to say this: You know those setups that some people will use where there's a splitter and only one output is connected, but the unused one isn't "turned off" with a filter (oh, and the days before there was such a feature...), or the side-loading to underground belts to only split one lane off a belt, and how both of these often leave items on them that will never be used? Yeah, those annoy me to no end.

[TheRailmaker]

Currently, construction bots try to find the chest that is the closest to the ghost they want to rebuild.

Idea is, that players want to bluprints to be built fast, so construction robots shouldn't travel over large distances (system that manages this is still pretty simple, though, so it's not that great) ... and you can use buffer chests to make logistic bots bring items closer to the construction site and construction bots will use the closest chest.

Lets make an example

if you deconstruct 1000 belts, and rebuild them later.
do you want the storage ones to be used first, or new 1000 belts from a closer chest ?

[Koub]

Currently, construction bots try to find the chest that is the closest to the ghost they want to rebuild.

Idea is, that players want to bluprints to be built fast, so construction robots shouldn't travel over large distances (system that manages this is still pretty simple, though, so it's not that great) ... and you can use buffer chests to make logistic bots bring items closer to the construction site and construction bots will use the closest chest.

Lets make an example

if you deconstruct 1000 belts, and rebuild them later.
do you want the storage ones to be used first, or new 1000 belts from a closer chest ?

I want to use the closest, for the fastest building time. I don't care if it's the storage or the provider chest. But I'll make sure I won't start reproducing belts before I've used up enough of the belts stored in my whole logistic network (through controlling the inserters on my belt factory).

[FuryoftheStars]

I'm still in the storage first camp. Even (now) knowing the logistic network is wireless, that's still config that has to be done to all involved inserters, either on the input or output side. Plus, unless you set the condition for something stupid low, or set up other clunky solutions, there will always be lots of the items left in your storage, but then there are side effects to these as well.

At least with storage first, they're only slow until storage is used up. Clunky solutions, otoh, will always be clunky.

[Nidan]

You can solves the problem by getting rid of passive providers: Storage chests can be filtered/restricted to one item, so if your passive providers only have one item you could replace them with filtered storage chests.

[FuryoftheStars]

You can solves the problem by getting rid of passive providers: Storage chests can be filtered/restricted to one item, so if your passive providers only have one item you could replace them with filtered storage chests.

Already been suggested. Point 4 on my summed up post back a page: viewtopic.php?p=569525#p569525

[Abarel]

if you deconstruct 1000 belts, and rebuild them later.
do you want the storage ones to be used first, or new 1000 belts from a closer chest ?

That's a tricky question, and I would like to answer "none of these options"; I chose a different way (see below), but if I need to chose between these 2, I prefer new belts from a closer chest.

On my ideal game, bots never fly long distances to pick materials to build something: there are materials near the construction site, usually delivered by train (and viceversa, bots place deconstructed materials into nearby chests, then delivered back to main storage by train). That works for whatever kind of chests you prefer. And the bot network at the construction place is isolated from the rest of the networks, so the bots just wait for a new train if more materials are needed.

[TheRailmaker]

On my ideal game, bots never fly long distances to pick materials to build something: there are materials near the construction site, usually delivered by train (and viceversa, bots place deconstructed materials into nearby chests, then delivered back to main storage by train). That works for whatever kind of chests you prefer. And the bot network at the construction place is isolated from the rest of the networks, so the bots just wait for a new train if more materials are needed.

I Agree on using small logistic areas in genral. But for early and mid game, a bigger area can give you good speed bost, while you dont care about performace or anything. especially in multiplayer, this phase is the most inderesting one.

[-Holo-]

This is a question I have been meaning to ask for a while, but I'd like the option to change the bots priority to take for example from storage chests first , then passive, etc.
Or anothr way , have requester/buffer chest have a tickbox with "Take from storage chests first"

[MentalHospital]

I think this should be an option. There is a lot of people who want it to prioritize storage (myself included), as well as the ones who want it to be the other way.

[1WheelDude]

Wow, I'm just learning that construction bots dont do this. I assumed that they pulled from storage like logistic bots. My mall is built with passive chests because I expected bots to pick up machines from storages that I either discarded or deconstructed. Should be a toggleable feature I think.

[astroshak]

You can connect the inserters filling those mall Passive Providers to the wireless logistic network, so the assembly machines only make when there isn’t enough in network as opposed to merely not enough in that chest. That would be to avoid overproduction.

[Eketek]

My solution is to place virtually all bot-based logistics in a small logistics zone, use a mod (Construction Probe) to obtain construction events and convert them into localized circuit-network signals, and a pile of combinators to use said signals to direct logistic bots at to load construction materials onto trains (each train dedicated to a specific logistic node). Since it is logistic bots doing all the picking, storage chests are implicitly prioritized for all construction which takes place outside of the central logistic zone.

[NineNine]

overproduction.

I'm sorry, I don't understand what this is.

[astroshak]

Overproduction: a set of boxes full of 48 stacks (each) of expensive machinery, when all you want on hand is 1 or 2 stacks.

Just in case that was a serious point and not a joke.