Is a pure modular logistics system viable?
Moderator: bobingabout
Is a pure modular logistics system viable?
I'm looking at that "Disable Roboports" checkmark, and I'm itching to enable it.
I really like the modular logistics system with having one thing for each purpose, but everything I can find about how robots work tells me it really won't work at all.
I'm mainly concerned about robots landing after finishing a job. As far as I understand a robot will forcefully charge at its target destination (a robochest in this case) and will not even consider any other charge point. This seems to make charge pads more or less entirely useless except for select points on the outer edge of your network that require long journeys and mid-trip charging.
has anyone played without vanilla (and upgraded vanilla ofc) roboports? Can it handle the same amount of bots? How did you design your network?
I really like the modular logistics system with having one thing for each purpose, but everything I can find about how robots work tells me it really won't work at all.
I'm mainly concerned about robots landing after finishing a job. As far as I understand a robot will forcefully charge at its target destination (a robochest in this case) and will not even consider any other charge point. This seems to make charge pads more or less entirely useless except for select points on the outer edge of your network that require long journeys and mid-trip charging.
has anyone played without vanilla (and upgraded vanilla ofc) roboports? Can it handle the same amount of bots? How did you design your network?
Re: Is a pure modular logistics system viable?
My experience on the subject is one year old, so it may have change, but i don't think in the way of being less viable.
I don't particularly remember a setting to disable the vanilla roboports but i didn't need them anyway.
One example of use for the recharge pad is when you have 10 station unloading iron ore in a row, and you use robots to drop it all at the begining of the main bus in a single area of requesting chest. The areas where robots will "end" their delivery is for ALL at the same place while pick up places are spread out; if you use 1 robotchest at every station and 1 near the requesting chest, and no recharge pad, what will end up over time is all the bots will be in the same robot chest and will bottleneck the recharge capacity of the robotchest. This happens in vanilla too, some roboports will have no bots , others be full.
To balance that in vanilla you would place many roboports near the end point, and a few at each station, with the modular you just have to place 1 robot chest near the end point, and lots of recharge pads. No need for robot chest at each station, the bots will end up leaving them anyway, you could place a few recharge pads along the way, just in case, but the regular path of a bot will be leaving the "drop" area, where you place the robotchest, then go pick up some material, in the range of the antenna, and bring it back near where it started, recharge and start over.
Another way I played with the modular system is by making "roads" for robots, you can use the logistic relay to shape the logistic network as very long and thin rectangle, or also you can make many tiny network close to each other and watch robots leave a chest, move an item, and get back in it. This way you can minimize the distance the bots will move, since each network is like 5x5 and has only 1 provider for 2 requester. There are many "new" possibilities.
But you can also keep your vanilla network ideas, and replace each roboport with a set of 1 robotchest, 1 recharge pad, and 1 logistic relay. Then realise that some chest are always empty so remove them, and some recharge pads are always full so double them. Since in vanilla sometimes we place roboports to extend the range (and we pay for the electricity of the buffer), sometimes to recharge more( quite often), sometimes to store more robots but it's usually charging that is a problem first.
So i would advise you to stop itching and disable vanilla roboports and just build more recharge pads if your concern turns out to be a problem
I don't particularly remember a setting to disable the vanilla roboports but i didn't need them anyway.
One example of use for the recharge pad is when you have 10 station unloading iron ore in a row, and you use robots to drop it all at the begining of the main bus in a single area of requesting chest. The areas where robots will "end" their delivery is for ALL at the same place while pick up places are spread out; if you use 1 robotchest at every station and 1 near the requesting chest, and no recharge pad, what will end up over time is all the bots will be in the same robot chest and will bottleneck the recharge capacity of the robotchest. This happens in vanilla too, some roboports will have no bots , others be full.
To balance that in vanilla you would place many roboports near the end point, and a few at each station, with the modular you just have to place 1 robot chest near the end point, and lots of recharge pads. No need for robot chest at each station, the bots will end up leaving them anyway, you could place a few recharge pads along the way, just in case, but the regular path of a bot will be leaving the "drop" area, where you place the robotchest, then go pick up some material, in the range of the antenna, and bring it back near where it started, recharge and start over.
Another way I played with the modular system is by making "roads" for robots, you can use the logistic relay to shape the logistic network as very long and thin rectangle, or also you can make many tiny network close to each other and watch robots leave a chest, move an item, and get back in it. This way you can minimize the distance the bots will move, since each network is like 5x5 and has only 1 provider for 2 requester. There are many "new" possibilities.
But you can also keep your vanilla network ideas, and replace each roboport with a set of 1 robotchest, 1 recharge pad, and 1 logistic relay. Then realise that some chest are always empty so remove them, and some recharge pads are always full so double them. Since in vanilla sometimes we place roboports to extend the range (and we pay for the electricity of the buffer), sometimes to recharge more( quite often), sometimes to store more robots but it's usually charging that is a problem first.
So i would advise you to stop itching and disable vanilla roboports and just build more recharge pads if your concern turns out to be a problem
Re: Is a pure modular logistics system viable?
My concern is that, according to anything I can find, robots going back to a chest after having made a delivery won't ever use a chargepad at all. Their destination has to be something they can land in, so they choose a robochest and then queue up for the one chargepoint, ignoring the chargepads right next to it.
Re: Is a pure modular logistics system viable?
You can try it for yourself first in a small new map using /cheat all, 1 robot chest, full of robots, and 0 or 50 chargepoint to see the difference.To my memory robots will consider recharge pads the closest near the robotchest they want to go if said robotchest is already charging the maximum number of robots it can charge at a time, making a chest with some chargepoint next to it act as a vanilla roboport with more charging capacity. It will not prevent robots charging on a rechargepoint to enter the robotchest if it is charging robots already, to my memory.
- bobingabout
- Smart Inserter
- Posts: 7352
- Joined: Fri May 09, 2014 1:01 pm
- Contact:
Re: Is a pure modular logistics system viable?
The problem with charging at the robochest is if they make it to the robochest and their charge level is below the minimum dock level, they WILL charge at the robochest.
however, if they drop to the minimum operating charge level while on the way to the robochest, they'll divert and charge at the nearest chargepad (or robochest if it's the closest)
That's basically the issue, there is no logic to which they use, they just want to use the closest, even if it has the biggest queue, and minimum number of pads.
however, if they drop to the minimum operating charge level while on the way to the robochest, they'll divert and charge at the nearest chargepad (or robochest if it's the closest)
That's basically the issue, there is no logic to which they use, they just want to use the closest, even if it has the biggest queue, and minimum number of pads.
Re: Is a pure modular logistics system viable?
During 0.16 I used this mod for recharging:
SmarterBotRecharge
Instead of forming a massive queue at a single roboport, you could place charging pads beside it which they'd also use to keep the logistics going. This was ideal for heavy traffic areas where a dozen charging pads could see constant use instead of just prioritizing only the nearest pad with a never ending queue.
Unfortunately it needs a 0.17 update but the bot efficiency and delivery times were night and day after installing it.
SmarterBotRecharge
Instead of forming a massive queue at a single roboport, you could place charging pads beside it which they'd also use to keep the logistics going. This was ideal for heavy traffic areas where a dozen charging pads could see constant use instead of just prioritizing only the nearest pad with a never ending queue.
Unfortunately it needs a 0.17 update but the bot efficiency and delivery times were night and day after installing it.
Re: Is a pure modular logistics system viable?
1) Why do robochest have a charging port at all?bobingabout wrote: ↑Wed Jul 03, 2019 12:14 am The problem with charging at the robochest is if they make it to the robochest and their charge level is below the minimum dock level, they WILL charge at the robochest.
however, if they drop to the minimum operating charge level while on the way to the robochest, they'll divert and charge at the nearest chargepad (or robochest if it's the closest)
That's basically the issue, there is no logic to which they use, they just want to use the closest, even if it has the biggest queue, and minimum number of pads.
2) There should be some game logic for bots to pick another near charging port if the queue is longer than the distance to the charging port. When a bot needs to charge it doesn't pick the closest but it adds some variance. So if you have 2x2 charging pads they will all be used to some extend. Is that logic failing for chests because the bots will already be so close to the chests port that picking any other would be too far away?
- bobingabout
- Smart Inserter
- Posts: 7352
- Joined: Fri May 09, 2014 1:01 pm
- Contact:
Re: Is a pure modular logistics system viable?
see point 1.
Before a recent update, if there was no chargepad on the robochest, the robot would come up to dock with it, then join the charge queue, waiting to charge at a non-existant charging pad.bobingabout wrote: ↑Wed Jul 03, 2019 12:14 am The problem with charging at the robochest is if they make it to the robochest and their charge level is below the minimum dock level, they WILL charge at the robochest.
Since a recent update, if you define a roboport with storage, but no chargepad, the game errors and won't load. it's not allowed.
I just use the base game logic, so, whatever that tells the robots to do, is what they do. the thing about robots in the base game is, to keep them as optimised as possible (use least CPU power) they're pretty dumb.mrvn wrote: ↑Wed Jul 03, 2019 9:34 am 2) There should be some game logic for bots to pick another near charging port if the queue is longer than the distance to the charging port. When a bot needs to charge it doesn't pick the closest but it adds some variance. So if you have 2x2 charging pads they will all be used to some extend. Is that logic failing for chests because the bots will already be so close to the chests port that picking any other would be too far away?
Re: Is a pure modular logistics system viable?
Well It may have changed since last time I used them, from what i am reading, i need to try those again before i give more advices !
Robots looking for the closest point to recharge when they run low, is fine, they will not clog the robotchest.
Robots STILL using the robotchest because they are low power but not enough for them to consider charging in the way are dumb.
Then you are the one that needs be smart and not allow them to run that low or force them to be low enough to charge in the way.
This means there is a non-optimal distance travel for robots which is long enough that they feel the need to charge at the chest and too short for them to charge in the way.
Now I am unsure but i guess their "dock level" is low enouh that it still allow to have a reasonable area around the chest where robots wouldn't charge after 1 trip I would probably design around small network that fits this range for high throughput so that the bots never charges on those, and would make it different for a mall or a construction area or a solar field, using charging pads all along.
Robots looking for the closest point to recharge when they run low, is fine, they will not clog the robotchest.
Robots STILL using the robotchest because they are low power but not enough for them to consider charging in the way are dumb.
Then you are the one that needs be smart and not allow them to run that low or force them to be low enough to charge in the way.
This means there is a non-optimal distance travel for robots which is long enough that they feel the need to charge at the chest and too short for them to charge in the way.
Now I am unsure but i guess their "dock level" is low enouh that it still allow to have a reasonable area around the chest where robots wouldn't charge after 1 trip I would probably design around small network that fits this range for high throughput so that the bots never charges on those, and would make it different for a mall or a construction area or a solar field, using charging pads all along.
- bobingabout
- Smart Inserter
- Posts: 7352
- Joined: Fri May 09, 2014 1:01 pm
- Contact:
Re: Is a pure modular logistics system viable?
Depends how busy you keep your robots too. if you keep them busy enough, they don't dock, they go from job to job, which keeps the robochests rather clear.mmmPI wrote: ↑Wed Jul 03, 2019 11:54 am Well It may have changed since last time I used them, from what i am reading, i need to try those again before i give more advices !
Robots looking for the closest point to recharge when they run low, is fine, they will not clog the robotchest.
Robots STILL using the robotchest because they are low power but not enough for them to consider charging in the way are dumb.
Then you are the one that needs be smart and not allow them to run that low or force them to be low enough to charge in the way.
This means there is a non-optimal distance travel for robots which is long enough that they feel the need to charge at the chest and too short for them to charge in the way.
Now I am unsure but i guess their "dock level" is low enouh that it still allow to have a reasonable area around the chest where robots wouldn't charge after 1 trip I would probably design around small network that fits this range for high throughput so that the bots never charges on those, and would make it different for a mall or a construction area or a solar field, using charging pads all along.
So, having too many robots on your network slows down efficiency, because they want to dock after each job (because there's always one waiting to pick up the job in storage)
Re: Is a pure modular logistics system viable?
Which also means the problem should solve itself.bobingabout wrote: ↑Wed Jul 03, 2019 12:08 pmDepends how busy you keep your robots too. if you keep them busy enough, they don't dock, they go from job to job, which keeps the robochests rather clear.mmmPI wrote: ↑Wed Jul 03, 2019 11:54 am Well It may have changed since last time I used them, from what i am reading, i need to try those again before i give more advices !
Robots looking for the closest point to recharge when they run low, is fine, they will not clog the robotchest.
Robots STILL using the robotchest because they are low power but not enough for them to consider charging in the way are dumb.
Then you are the one that needs be smart and not allow them to run that low or force them to be low enough to charge in the way.
This means there is a non-optimal distance travel for robots which is long enough that they feel the need to charge at the chest and too short for them to charge in the way.
Now I am unsure but i guess their "dock level" is low enouh that it still allow to have a reasonable area around the chest where robots wouldn't charge after 1 trip I would probably design around small network that fits this range for high throughput so that the bots never charges on those, and would make it different for a mall or a construction area or a solar field, using charging pads all along.
So, having too many robots on your network slows down efficiency, because they want to dock after each job (because there's always one waiting to pick up the job in storage)
- Robots fly away for a job.
- They come back and get stuck in the charging queue.
- Fewer robots are available to do work.
- Robots will be re-tasked before they reach the chest.
- Robots charge on the fly and not at the chest.
Now if you send out 5000 construction bots and they get stuck at chest and then you want to clear another 5000 trees directly after that is bad. Takes forever to get them free again. But logistic bots are generally more a steady stream.
Unless, from when I tried it, when you use bots for train unloading. Then every time a train leaves all the bots will go to recharge at their chests. And my experience from maybe 2 years ago was that with chests they recharge one by one but with reboports they simply all go into the roboport. Different minimum charge level?
Re: Is a pure modular logistics system viable?
It is a poor advice to actually disabled vanilla roboport completly then, you could test both setups and make some obstacle courses for the robots to see what distance they can do without the need to charge with classic roboports and compare that with robotchest to define the size of the areas you can build on.
And if you make mistakes you can just research faster or better bots, at some point they are so fast and carry so much that you really need to move millions of material to use a lot of them simultaneously.
And if you make mistakes you can just research faster or better bots, at some point they are so fast and carry so much that you really need to move millions of material to use a lot of them simultaneously.
Re: Is a pure modular logistics system viable?
That's what I thought... I want it to work, but the game engine in its current state just makes it too much of a hassle.It is a poor advice to actually disabled vanilla roboport completly then
Maybe we can convince the devs to have queued robots search for a new charging port at a certain queue length (just scan the surrounding 8 chunks, no more), that shouldn't be too much of a performance drain, since bots do that anyway when searching for a landing place.
Re: Is a pure modular logistics system viable?
Well you could play without bots in vanilla at all and still send the rocket , or build a megabase. So you could do the same with the modular roboports. Not using any and still enjoy your base; if you WANT IT to work, you can MAKE IT WORK.
I wouldn't want to ask for something before i have tried everything i could and it still doesn't work. But even in this case i'm pretty sure some players would have found work around what i couldn't do.
It makes it harder not impossible :p
I wouldn't want to ask for something before i have tried everything i could and it still doesn't work. But even in this case i'm pretty sure some players would have found work around what i couldn't do.
It makes it harder not impossible :p
- bobingabout
- Smart Inserter
- Posts: 7352
- Joined: Fri May 09, 2014 1:01 pm
- Contact:
Re: Is a pure modular logistics system viable?
heck... I'd say "same chunk" but since you can't see chunks, you could end up building on a chunk boundary, so... within 32 tiles (the size of a chunk)jinks wrote: ↑Wed Jul 03, 2019 4:52 pm Maybe we can convince the devs to have queued robots search for a new charging port at a certain queue length (just scan the surrounding 8 chunks, no more), that shouldn't be too much of a performance drain, since bots do that anyway when searching for a landing place.
Re: Is a pure modular logistics system viable?
After further experiment I discovered that robotchest are greatly overated, you don't need to place them at all just some charge pads and having the robots flying 100% of time is fine. When they don't know where to go they just stand there instead of in a chest and that's just fine.
Good thing i have much free time.
If you were to place robotchest, It seems to be the best when they are not in the way of the robots, that's a bit weird but this way the robots never tries to recharge in them, just on the recharge pads that are in their path between receiver and provider chest.
Now away is even better when it's aligned in this order:
Provider chest - Recharge pad - recharge pad - recharge pad - Receiver chest - recharge pad - Robot chest
This way most of the robots don't go past the receiver chest, and when they do they often recharge in the pad, and you end up with the excess robots in the chest.
Or in this order :
Robot chest -Recharge pad - Provider chest - Recharge pad - recharge pad - recharge pad - Receiver chest
This way all the robots start their way back as soon as they deliver something, aiming for the robot chest, charging in the way, and being rerouted before they reach the chest.
Using two robot chest in those conditions is not optimal.
Here is a save with 10 stations with different placement for the elements. If you want to see for yourself. There are only 2 here but the saves has all 10. On the left is only 1 chest and 0 recharge pads
On the right 1 chest and Recharge pads everywhere.
On the left side you can see many robots charging in the pads, that was at the opening of the station when a cloud of robots went all at once.
On the right side is 300 bots roughly unloading 4 belt of iron ore, without any robotchest in their network, they can never rest .
Good thing i have much free time.
If you were to place robotchest, It seems to be the best when they are not in the way of the robots, that's a bit weird but this way the robots never tries to recharge in them, just on the recharge pads that are in their path between receiver and provider chest.
Now away is even better when it's aligned in this order:
Provider chest - Recharge pad - recharge pad - recharge pad - Receiver chest - recharge pad - Robot chest
This way most of the robots don't go past the receiver chest, and when they do they often recharge in the pad, and you end up with the excess robots in the chest.
Or in this order :
Robot chest -Recharge pad - Provider chest - Recharge pad - recharge pad - recharge pad - Receiver chest
This way all the robots start their way back as soon as they deliver something, aiming for the robot chest, charging in the way, and being rerouted before they reach the chest.
Using two robot chest in those conditions is not optimal.
Here is a save with 10 stations with different placement for the elements. If you want to see for yourself. There are only 2 here but the saves has all 10. On the left is only 1 chest and 0 recharge pads
On the right 1 chest and Recharge pads everywhere.
On the left side you can see many robots charging in the pads, that was at the opening of the station when a cloud of robots went all at once.
On the right side is 300 bots roughly unloading 4 belt of iron ore, without any robotchest in their network, they can never rest .
- bobingabout
- Smart Inserter
- Posts: 7352
- Joined: Fri May 09, 2014 1:01 pm
- Contact:
Re: Is a pure modular logistics system viable?
The down side of no chests is on really large networks, or networks with a large number of robots. having 75 thousand robots constantly in the air isn't very UPS friendly.
I had that many, I cut it down to only 25 thousand.
I had that many, I cut it down to only 25 thousand.
Re: Is a pure modular logistics system viable?
Good tip about putting the robochest out of the flight path.
When bots run low on power they pick an entity for recharge. They do not care if the entity has 1, 4 or 9 recharge ports. But obviously a 9 port entity is much better in handling a queue of bots. So it makes sense to hide the 1 port chest behind a 4/9 port recharge pad.
As for having no chests: Bots that hang around use up power I think. Can you confirm that the idle bots will go for a recharge after a while?
When bots run low on power they pick an entity for recharge. They do not care if the entity has 1, 4 or 9 recharge ports. But obviously a 9 port entity is much better in handling a queue of bots. So it makes sense to hide the 1 port chest behind a 4/9 port recharge pad.
As for having no chests: Bots that hang around use up power I think. Can you confirm that the idle bots will go for a recharge after a while?
Re: Is a pure modular logistics system viable?
They do recharge, my idea is to have just the correct number or a little less than what is needed for the task in an isolated network. This way you can easily cover the path with recharge pads that they will use for very short time, as soon as they level drop a threshold, a flying belt in a way.
You can first put the chest away from the path, and when it stabilize you pick it up entirely with the excess robots. Then just maintain the flow or they will look weird waiting in the air. I didn't tested that in particular but in those cases it seemed that they stay where they are when they stopped having a thing to do until they reach a charge threshold, then go recharge to the closest entity ( or normal behavior when charge drops) they stay there.
The place they stop is very often a receiver chest i guess but that could be a problem for robots that repair defense, or just build.
You can first put the chest away from the path, and when it stabilize you pick it up entirely with the excess robots. Then just maintain the flow or they will look weird waiting in the air. I didn't tested that in particular but in those cases it seemed that they stay where they are when they stopped having a thing to do until they reach a charge threshold, then go recharge to the closest entity ( or normal behavior when charge drops) they stay there.
The place they stop is very often a receiver chest i guess but that could be a problem for robots that repair defense, or just build.
Re: Is a pure modular logistics system viable?
If you use the bots for the factory then having many separate logistic nets is a good tip too. Covering your whole megabase with just one big logistic zone doesn't work well.mmmPI wrote: ↑Thu Jul 04, 2019 1:58 pm They do recharge, my idea is to have just the correct number or a little less than what is needed for the task in an isolated network. This way you can easily cover the path with recharge pads that they will use for very short time, as soon as they level drop a threshold, a flying belt in a way.
You can first put the chest away from the path, and when it stabilize you pick it up entirely with the excess robots. Then just maintain the flow or they will look weird waiting in the air. I didn't tested that in particular but in those cases it seemed that they stay where they are when they stopped having a thing to do until they reach a charge threshold, then go recharge to the closest entity ( or normal behavior when charge drops) they stay there.
The place they stop is very often a receiver chest i guess but that could be a problem for robots that repair defense, or just build.