Supply Network Management

[mcwaffles2003]

I wanted to make a thread specifically for this issue as I've seen the conversation several times in different threads.

So, should automation be part of this game? If so, to what extent and in what manner?

I think it's fair to say that most of us know the pain that can come in the mid/late game. You have a station or two, 20+ communication satellites, 3 or 4 active probes going to different locations, or whatever it may be. Regardless, keeping up with all of these different things going on continuously and keeping them maintained can make the late game turn dull or boring ("milk runs"). I feel if this issue isn't tackled head on managing the normal drag from KSP 1 with the addition of running multiple colonies and continuously supplying multiple orbital shipyards stacked on top then essentially no ones ever going to get interstellar anyways.

So far I see 3 levels of automation:

1. None - The game continues, as is, and every maneuver is handeled manually
2. Minimal - Rockets are made as normal, launched to orbit, then a series of maneuvers are planned, and are performed in the background as their resources are taken into account and staged accordingly. Actions such as docking, take off, and landing are performed manually still
3. Some - Similar to minimal but take off, landing, and docking are also planed ahead of time and automated as well.
4. Lots - Resources for each colony/station can be monitored and "milk runs" can be ordered. Either default transport craft (depending on tech level) or user built craft can have entire routes simulated and effectively a run is ordered and all that matters is the time to wait for the resources to arrive.

Obviously, these are not the only ways this problem can be solved. Please comment your own solutions/opinions of the solutions 

Edited February 13, 2020 by mcwaffles2003

[Guest]

As I said in the other thread I think option 1 to 3 has nothing to do with the problem and 4 includes dozens of options and different systems, from recording a mission you've already done to implementing a fully fledged NPC ships and companies system with a whole new layer of Gameplay.

 

[DStaal]

Are we talking about automation, or about supply networks?  Automation is one way to set up supply networks, but it's not the only way, and it's not the only thing automation can be used for.  They are really separate discussions.

[mcwaffles2003]

6 minutes ago, Master39 said:

As I said in the other thread I think option 1 to 3 has nothing to do with the problem and 4 includes dozens of options and different systems

I hear you and that's why I made the 4th option, if you feel your idea wasn't represented correctly please correct me so I can amend it, and if you'd like to flesh out the idea more or expand on any other systems please do so. I just would like to see where community consensus currently lies and to what degree it is comfortable with automation

 

2 minutes ago, Master39 said:

 from recording a mission you've already done

I've thought about this one but wouldnt that require identical starting/ending points to properly simulate? for instance, record the mission as launched from the optimal transfer window then request an automated one from the least optimal transfer window. Perhaps the ship built for the 1st example doesn't even have enough dV to even handle the 2nd example. 

 

Just now, DStaal said:

Are we talking about automation, or about supply networks?  Automation is one way to set up supply networks, but it's not the only way, and it's not the only thing automation can be used for.  They are really separate discussions.

Fair point, will change the thread title. Point of thread is supply networks and avoiding mid-late game drag

[Dirkidirk]

i am 82% sure that you work for star theory, and are looking for the best form of Supply Network Management for the game, that the most people would lIKE 

Edited February 13, 2020 by Dirkidirk
give us info!!!!

[mcwaffles2003]

5 minutes ago, Dirkidirk said:

i am 82% sure that you work for star theory, and are looking for the best form of Supply Network Management for the game, that the most people would lIKE 

Nah, but I make a lot of threads with polls so ST can get quantifiable data on what the community wants. If they're not gunna actively participate in the convo I'm hoping they're at least watching the conversations... and if that's the case I want them to be at their best ability to serve the community theyre working for

[DStaal]

For supply networks, I honestly feel something like Pathfinder has is probably best: There is a specific part that you need to have at each nodepoint, that requires resources to run and to send/receive transports.  A formula is worked out that calculates cost of transfers, which is based on a semi-pessimistic scenario, for both time and resources used.  There's no physical presence to the transfer mission.

You then have two choices: Get the parts into place, with the resources they need to run (Pathfinder actually requires constant research points - which could also be converted to science...) and set up transfers at a high cost, or do the transfers yourself at a lower cost in terms of resources and/or time.  Extend that with a 'repeat every X days' or a 'repeat when X is lower than Y on ship/base/planet A', and you have a fairly hands-off network that requires setup to get going, without much micromanagement - but the micromanagement might still be worth it in certain situations.

[Terwin]

Star Theory will need something simple and robust or else their colonization will be dead in the water.

I expect something along the lines of 'Task: set up a water-supply   Reward: colony size cap raised from 10 to 50.'

where the details are like one of 'Set up a facility able to refine X units of water/day' OR 'Build a facility on this planet with a Drill, water refinery, and pumping station'  Possibly with a maximum range from the colony.

 

This requires at-most colony-based missions.  No background processing, no programming languages, no dv calculations, no on-going logistics, etc.  It also meshes well with the Event based colony growth and 'some form of life-support' which have both already been announced.  (This is basically the sort of thing I have been expecting since hearing about the 'some form of' life support and 'event based colony growth.' )

Anything with on-going logistics will also be a non-starter for interstellar flights because the time-warp to get to another star will cause anything with a chance of failure from neglect to fail.

If interstellar flight requires an off-kerbin base, then that base will need to be self-sufficient over long time-spans if any interstellar flight is to ever arrive.

 

I would expect orbital stations to have missions like 'designate a ground-base with at least X population and a launch facility to increase population cap from 5 to 10' or 'Enlarge designated ground base to produce X of resource A, Y of Resource B, and Z of resource C to allow assembly of a ship construction dock'

 

Simple, easy to understand, easy to ignore before or after, and most importantly, no gotchas for inexperienced players.

 

Even RoverDude's excellent MKS planetary warehouses functionality is too complex for inclusion in the base game because KSP is not a logistics game, KSP is a game about building and flying rocket-ships and accidentally learning about space as a by-product.  Keep the game-focus simple, and put focus-changing modifications in mods and DLC for those that want it.

[mcwaffles2003]

2 hours ago, DStaal said:

For supply networks, I honestly feel something like Pathfinder has is probably best: There is a specific part that you need to have at each nodepoint, that requires resources to run and to send/receive transports.  A formula is worked out that calculates cost of transfers, which is based on a semi-pessimistic scenario, for both time and resources used.  There's no physical presence to the transfer mission.

You then have two choices: Get the parts into place, with the resources they need to run (Pathfinder actually requires constant research points - which could also be converted to science...) and set up transfers at a high cost, or do the transfers yourself at a lower cost in terms of resources and/or time.  Extend that with a 'repeat every X days' or a 'repeat when X is lower than Y on ship/base/planet A', and you have a fairly hands-off network that requires setup to get going, without much micromanagement - but the micromanagement might still be worth it in certain situations.

I have never used pathfinder, but isn't it essentially a heavy reliance on ISRU?

[Guest]

Yeah I'm with @Master39 and @DStaal. Automation is one thing and I would definitely like to see it -- for example, high-level probe cores that handle certain manoeuvres on their own, and can be programmed with a scripting language. 

But that really doesn't address the milk run problem; Option 4 is the only way to go with that -- the specifics of it can vary, and it is not going to be trivial to do in a way that's engaging and fun. But I can't see how the late game could work without it; other than by removing or trivialising resource management so that milk runs aren't necessary, which would be a bit of a let-down.

[DStaal]

42 minutes ago, mcwaffles2003 said:

I have never used pathfinder, but isn't it essentially a heavy reliance on ISRU?

Like MKS, it's a colonization mod, so yes there's a lot of ISRU.  It's real emphasis is to be able to bootstrap yourself up from a fairly small initial landing into a full colony.  There's lots of features, for this discussion I'm focusing on it's 'Pipeline' parts - which are there to help you move resources around, in case you don't have everything you need in one place.  (Or even on one planet.)  Which is pretty common...

IIRC, it only works within a single SOI - but it can do the ground-to-orbit link, so you can then transfer from there.  It's not a complete solution to the problem, but I think it's a decent model for a starting point.

[mcwaffles2003]

27 minutes ago, Brikoleur said:

Yeah I'm with @Master39 and @DStaal. Automation is one thing and I would definitely like to see it -- for example, high-level probe cores that handle certain manoeuvres on their own, and can be programmed with a scripting language. 

But that really doesn't address the milk run problem; Option 4 is the only way to go with that -- the specifics of it can vary, and it is not going to be trivial to do in a way that's engaging and fun. But I can't see how the late game could work without it; other than by removing or trivialising resource management so that milk runs aren't necessary, which would be a bit of a let-down.

The entire idea behind resources in mid-late game seems a bit confusing personally. I don't see it being hard to have mining colonies producing an excess of a rare mineral and selling it off in some way to make essentially infinite wealth. I'm wondering if the career mode will start as money dependent to build craft and move on to a form of resource dependence instead of money dependence.

Personally, I'm hoping that once sufficient colonization has begun, there will be some form of window/new KSC building wherein there is an easy to discern state of resources for any and all outposts/colonies

From this window one could order resource transfers between outposts/colonies based on their ability to:

• Build a ship capable of making the transit (Its own VAB/BAE)
• Launch a ship (Its own launch pad)
• The colony/outpost HAS the requested resources (To make the ship and fulfill the request)
• Destination has the ability to receive the requested goods (docking port)

From this a transfer can be started and maybe in the tracking station the transfer can be monitored, though not piloted.

Assuming this kind of system would be favorable I'm curious about anyone position on whether the transfer ship should be built by the player or simply generated and never seen. If it is built by the player, perhaps it needs to be demonstrated, via a single manual run, that the ship can:

• Take off (seperate certification upon making it through an atmosphere)
• Survive transit
• And IMO most importantly, survive a re-entry to make the delivery.

This may be more involved, possibly defeating the point of creating automated transport to a degree, but keeps a form of realism as a ship without a heat shield 

Another idea could be to break transport into stages, where transport is based on orbital distribution hubs where automated shipping to a ground colony wanting to receive shipments must have a corresponding low orbit hub that can supply easy package drops to the colonies dependent upon them.

 

Any comments, criticisms, opinions, improvements on any of this? Is it even an agreeable start?

22 minutes ago, DStaal said:

Like MKS, it's a colonization mod, so yes there's a lot of ISRU.  It's real emphasis is to be able to bootstrap yourself up from a fairly small initial landing into a full colony.  There's lots of features, for this discussion I'm focusing on it's 'Pipeline' parts - which are there to help you move resources around, in case you don't have everything you need in one place.  (Or even on one planet.)  Which is pretty common...

IIRC, it only works within a single SOI - but it can do the ground-to-orbit link, so you can then transfer from there.  It's not a complete solution to the problem, but I think it's a decent model for a starting point.

I'm hoping a large component to colonies is ISRU dependent, at least then need for logistics would be  minimized

Edited February 13, 2020 by mcwaffles2003

[shdwlrd]

It would be necessary to have some automation for the late game resource transfers.

They way I see it you would have to build or place specific parts on your ground/orbital colonies to use a connection point to automate the resource distribution network. For ground based colonies, a landing pad or hangar/garage. For orbital colonies a dedicated docking port, or hangar. Once included in the network, you can see the resources flowing in. 

The different options could include building your own crafts, setting the rules how the resources are transferred (time intervals, how the resources are routed, etc.), have transfers happen completely automagically, paying resource/funds/time costs for transportation of resources by private contractors, whether or not to have random hiccups. (Just spitballing options, I’m not really thinking about the details here.)
 

[DStaal]

Thinking about it further - I actually don't want this to be too automated.  Transferring resources around would be a real 'pain point' in interplanetary or interstellar colonization, and it would lose it's impact if it's automated to a background task where you can click a few buttons to set up a transfer when your snacks get low.  You definitely want to automate it a little so not every transfer has to be done by hand, but where is the line drawn?

I play Offworld Trading Company on occasion as well, and the fact that it's set on a different planet is really inconsequential to the game: It takes a certain amount of time to ship goods to/from market, but how far away that market is doesn't really matter.  KSP needs to be more realistic than that.  If I can just set up a transfer to Duna - or even a recurring transfer to Duna - of supplies, what does it matter that this is space sim?  (Even if I have to fly it the first time.)

I'm leaning more towards any transfers should be limited to your current SOI, and likely only to designated orbital platforms (surface I'd be a bit more fluid about).  Would fleets of transport ships between planets make sense?  (Automated or not?)

[mcwaffles2003]

1 minute ago, DStaal said:

Thinking about it further - I actually don't want this to be too automated.  Transferring resources around would be a real 'pain point' in interplanetary or interstellar colonization, and it would lose it's impact if it's automated to a background task where you can click a few buttons to set up a transfer when your snacks get low.  You definitely want to automate it a little so not every transfer has to be done by hand, but where is the line drawn?

I play Offworld Trading Company on occasion as well, and the fact that it's set on a different planet is really inconsequential to the game: It takes a certain amount of time to ship goods to/from market, but how far away that market is doesn't really matter.  KSP needs to be more realistic than that.  If I can just set up a transfer to Duna - or even a recurring transfer to Duna - of supplies, what does it matter that this is space sim?  (Even if I have to fly it the first time.)

I'm leaning more towards any transfers should be limited to your current SOI, and likely only to designated orbital platforms (surface I'd be a bit more fluid about).  Would fleets of transport ships between planets make sense?  (Automated or not?)

Then what's your opinion on having the condition that each trade route has to be flown at least once manually to activate it and trade routes can only exist between orbital stations and not to ground colonies?

[linuxgurugamer]

There are several mods which implement this:

• Routine Mission Manager
• Davon Supply Mod
• Kerbal Space Transport System

are three that I know of.

Edited February 14, 2020 by linuxgurugamer

[shdwlrd]

3 hours ago, linuxgurugamer said:

Kerbal Space Transport System

I may be wrong, but wasn't Kerbal Space Transport System a parts pack that helped you land rovers and provided large cages to help launch components?

[shdwlrd]

5 hours ago, DStaal said:

I'm leaning more towards any transfers should be limited to your current SOI, and likely only to designated orbital platforms (surface I'd be a bit more fluid about).  Would fleets of transport ships between planets make sense?  (Automated or not?)

It makes sense limit ground to orbit or vice versa transfers to the current SOI. (If you really need something somewhere not within your current SOI ASAP, you're going to do that mission yourself.)

For station to station transfers the number of ships would depend on how much funds/resources do you want to spend on just to ferrying stuff around. For this to work, it will have to cost you something; whether it's funds, resources, or both. As an example, paying 90-95% of what you produce in fuel just to run a large transport fleet is a waste of time. It would force the player to choose the appropriate number of ships to dedicate to ferrying resources and Kerbals around. 

[linuxgurugamer]

1 hour ago, shdwlrd said:

I may be wrong, but wasn't Kerbal Space Transport System a parts pack that helped you land rovers and provided large cages to help launch components?

Nope.  I recently adopted it, no parts in it at all

[shdwlrd]

29 minutes ago, linuxgurugamer said:

Nope.  I recently adopted it, no parts in it at all

Huh, ok. You would know better than I would. Now I'm curious what that mod was. I'll have to dig it up later. 

[MechBFP]

The devs have already said that one of their primary goals is to allow people to ignore their colony’s and do whatever they want, which means free use of time warp without consequences.

So there will be zero need to do any manual supply runs, otherwise that would break their design goal. 

[DStaal]

42 minutes ago, MechBFP said:

The devs have already said that one of their primary goals is to allow people to ignore their colony’s and do whatever they want, which means free use of time warp without consequences.

So there will be zero need to do any manual supply runs, otherwise that would break their design goal. 

Sure, but there's multiple ways to accomplish that.  One way is to have supply routes you can set up in the background.  Another would be to allow the colonies to be self-sufficient.  A third would be for colonies to 'hibernate' under certain conditions - they don't produce anything, but they can still be productive if brought back to the foreground.  I'm sure there's others I'm not thinking of.

Maybe they can always be ignored.  Maybe they can be ignored under certain circumstances.  Maybe a full-size colony can ignore them, but you'll need them in some mid-stages.  Maybe Kerbonauts are immune to needing life support (but colonists - who we don't interact with - aren't). Etc.

I honestly would like supply runs to be useful, at least some of the time.

15 hours ago, mcwaffles2003 said:

Then what's your opinion on having the condition that each trade route has to be flown at least once manually to activate it and trade routes can only exist between orbital stations and not to ground colonies?

Depends on how it's modeled, I think, and how it's set up in the game.  I don't necessarily mind say railgun-based delivery networks, which you wouldn't fly the route first.  (You just have to have the right parts on each end, and the ends within a certain range.)  I'll admit I've never actually used Routine Mission Manager, which can do the same in KSP1.  It would definitely be more flexible.

There definitely need to be limits on any system.  I'm just not sure where I'd want to put the line between 'don't bore the player' and 'show the real complexity/cost of such a system'.

[mattinoz]

I think it's a chance to tutor the player to help them develop in the game and get beyond the Kerbin SOI. Conbined with much needed planning tools it could show the players the clockwork of how the system works. Help them learn how to hit targets cheap and slow or fast and expensive.

To me I'd think we could have missions on repeat that could be planned out and the game tell us then ho often it could be repeated and system would take care of dropping the craft in to timeline. I'd like to see stock as a simple resource (say tonnage) that could be converted as needed to other resources or at least the ones available at KSC. 

Player would need to fly first time then it just repeats as requested. 

Number of automated or tracked flights and the skills could be tied to the tracking centre level or better still staffing or probe core levels. 

Leave lots of room for mods to add complexity to suit the players level. of management. 

Edited February 15, 2020 by mattinoz

[Pthigrivi]

The simplest solution is to rely on IRSU and the ability to make 100% self sustaining bases so you didn't need supply runs. 

[Superfluous J]

I want some of 2 and some of 4 but none of 3.