What would a switch to n-body physics take?

[Glaran K'erman]

I know that it would take A LOT of computing power to switch to n-body physics but just how much, and on that, what are the other challenges?

[Superfluous J]

This has been asked many times before, and discussed to Moho and back.

[r4pt0r]

I think it would be bad for gameplay. orbital decay killing a different mission while you timewarp to jool would suck

[Robotengineer]

N-body physics would be like LS. It would require you to do station keeping on almost all vessels, which would be come a massive pain very quick.

[cantab]

"Station keeping" could easily be abstracted away, handled automatically in return for a slow drain of fuel. It would give satellites and similar things finite lifespans and encourage the player to make tradeoffs - a light satellite with little fuel that's cheap to launch but will soon lose stationkeeping, or a heavier satellite with extra fuel that's more expensive to launch but will maintain its desired orbit for longer.

[Vanamonde]

Squad has said that they have no desire to try to make a transition to N-body. In fact, it's on the What Not to Suggest List. Assuming that you just want to discuss it as an intellectual problem, though, this thread has been moved to the off-topic lounge. Keep in mind, though, that trying to persuade Squad to make the change will get the thread closed.

[PB666]

Squad has said that they have no desire to try to make a transition to N-body. In fact, it's on the What Not to Suggest List. Assuming that you just want to discuss it as an intellectual problem, though, this thread has been moved to the off-topic lounge. Keep in mind, though, that trying to persuade Squad to make the change will get the thread closed.

You mean to say if I want any thread closed all I need to do is write "Squad can you please may the SOI realistic". Wow thats like open saysame.

[pincushionman]

I know that it would take A LOT of computing power to switch to n-body physics but just how much' date=' and on that, what are the other challenges?[/quote']

Well, you can go over to the thread and discuss this with the guys that are actually trying to do it?

Hint: performance isn't that much worse than the current system. Navigation is somewhat more complicated, though.

[Flymetothemun]

A 1.21 Gigawatt Computer Power Supply

[Superfluous J]

You mean to say if I want any thread closed to get infractions and my posts deleted all I need to do is write "Squad can you please may the SOI realistic". Wow thats like open saysame.

There, I fixed that for you

[Teilnehmer]

A far as I understand, in patched conics, timewarp has a pretty little effect on accuracy. You can skip a collision at a high timewarp value, but the coordinates are functions of orbital parameters and time only and can be evaluated accurately regardless of timewarp.

In n-body physics, the coordinates of bodies depend on their previous values. During timewarp, you have to increase the distance between time samples so the accuracy degrades.

So timewarp becomes risky in n-body physics. While orbiting nicely at 1×, your satellite can suddenly fly away or collide with the planet when you switch to 1000×.

Please correct me if I’m not right.

Edited July 7, 2015 by Teilnehmer

[Glaran K'erman]

Squad has said that they have no desire to try to make a transition to N-body. In fact, it's on the What Not to Suggest List. Assuming that you just want to discuss it as an intellectual problem, though, this thread has been moved to the off-topic lounge. Keep in mind, though, that trying to persuade Squad to make the change will get the thread closed.

Yea lol I've seen the list, just wondering exactly the difficulties in making such a change would present. Obviously there are many as I have learned, thanks guys. Never thought too much about them incorporating stationkeeping (never knew it was called that but makes sense) but I would welcome the change and that wouldn't be to hard to simulate it seems.

Edit:

Whoa, ever heard of Principia? https://github.com/mockingbirdnest/Principia

Edited July 7, 2015 by Glaran K'erman

[Friend Bear]

------------------------------------------------------------------

I apologize if this is already discussed. However, I looked everywhere and I only found the link I am putting in below. No discussion has occurred recently that I could see.

------------------------------------------------------------------

What is the status of N-Body physics and its applied orbital characteristics in Kerbal Space Program, whether it be mod, core game development, or an interest, generally?

This discussion is meant to do two things:

(1) Determine the future of N-Body physics in KSP (whether prospective or set in stone) Without begging or demanding developers or others commit to this (e.g. - possibility of application, memory usage, applicability, etc.);

(2) Spur discussion of the topic of N-Body physics so as to interest developers and/or modders into pursuing its application.

As a jumping-off point, I found this thread and will read it further, but hopefully people here can discuss it if they have used it:

http://forum.kerbalspaceprogram.com/threads/68502-WIP-Principia-N-Body-Gravitation-and-Better-Integrators-for-Kerbal-Space-Program

[r4pt0r]

I like the SOI system as it makes orbits predictable. I dont need to worry about orbital decay or something stupid killing a ship while I am timewarping a different ship to jool for 3 years.

[Friend Bear]

I definitely agree and I will always love a game where that is in play, especially when you are just getting into the game because there is already so much to learn.

But, when you get 1,000 hours + of game time in, alternative modes of gameplay keeps things spicey.

[mikegarrison]

I like the SOI system as it makes orbits predictable. I dont need to worry about orbital decay or something stupid killing a ship while I am timewarping a different ship to jool for 3 years.

The various krakenspawn fulfill that ecological niche anyway. I've had many mysterious things kill ships without even having to introduce the holy mysteries of orbital permutations.

[Friend Bear]

*for some reason I can't quote things on this laptop*

I guess one major thing they need to address is the kraken. They have made huge imrpovements in that front, but at the same time, I wonder if code could be used to treat a structure as one thing rather than the individual components. Given, I know some VB and C++ and am probably way off, but once addressed more in depth I don't see why not.

[Motokid600]

I wouldn't mind an Nbody system that applied its physics to specified vessels. Like in the tracking station you have different types of vessels. I would have it only so debris and the active ship receive nbody calculations.

And if I wanted a hardcore mode id enabled nbody for all vessels.

[technicalfool]

Hi.

N-body physics probably isn't going to make it into the stock game. The Principia mod does implement N-body physics over KSP's stock orbital mechanics. As a WIP mod, it comes with no guarantees that it will not crash, corrupt saves or do various other interesting things. However, if you want wobbly orbits, L-points and moons getting flung out of the Jool system and into deep space, that's the one to try.

You can get the mod by joining the #principia channel on Espernet and asking. Be aware that the ops there may not always be awake, so hang around until you get an answer.

Edited July 15, 2015 by technicalfool

[Eric S]

While I'd like to play with N-body physics, I have to concur with r4pt0r in that it doesn't belong in the base game. I look forward to seeing Principia grow and become more stable and accessible.

[Vanamonde]

Friend Bear, Squad has no interest in adapting the game to use N-body simulation, under the feeling that it would complicate the software immensely without adding anything to gameplay. In fact, N-body simulation is on the forum's list of things not to suggest again, because it's been talked to death so much already. But for those interested in just talking about it under the understanding that it will not be in the game, your thread has been merged into the master thread for this subject.

[Friend Bear]

Exactly why my opening was worded the way it was. I never expected it to be nor did I ever ask or suggest it be.

I also mentioned mods.

This is purely a discussion of its use and the possibility of its implementation, core game or mod alike. Now that people have so strongly stated it should not be implemented in the core game, as I ultimately agree, lets talk about the possibility of mods implementing it or those that already have and their reliability.

----

I actually only included "core game" as a talking point. I think it is wholly far fetched to expect n-body in the core game. But the prospect of a highly difficult gravity mechanism is exciting, so the post is meant to capture all opinions and approaches. This is primarily because of my lacking familiarity with a lot of modern computer programming languages and my only elementary understanding of gravity and its complex, realistic application in an orbit scheme.

Edited July 16, 2015 by Friend Bear

[Superfluous J]

This is purely a discussion of its use and the possibility of its implementation, core game or mod alike. Now that people have so strongly stated it should not be implemented in the core game, as I ultimately agree, lets talk about the possibility of mods implementing it or those that already have and their reliability.

Have you tried Principia? Or do you not find that mod sufficient in some way?

Edited July 16, 2015 by 5thHorseman

[Friend Bear]

I actually haven't had time to try any yet so I am poking around. I will take a look. It seems way promising.

I also never considered that an orbital model would harm current planetary orbit in KSP (either because of their huge mass or because they are locked in railed orbits), so I am really considering all approaches before I dive in and spend a huge amount of time on a game where all planets end up crashing =P

I also wouldn't mind learning as much as I can about it so I have been researching. Principia's application seems accurate, but I also feel like the other bodies' gravitational pull on orbiting bodies, whether a satellite or otherwise, seems a little strong for the distance they are from that object.