Oh it's on, Nuclear Engineering, it's on.

  • Right, so as much as I love sitting here messing with the planner trying to achieve an extra 1 effective EU/t, and that there's magic to discovering things manually, I'm just a smidgen curious as to the possible configurations. So, I'm gonna write a program to do it for me. Heck, I'll even do it in java just for kicks. Heck, I might even give out the source code and post it here when I'm done or have a working prototype. Heck, I might even start right now... What I'd like to know, is A. Has it been done, and B. Does anyone care? See you in a few days.

  • Given each spot in the reactor may contain:


    [Empty (or empty like),Ice,FullWaterBucket,lavabucket,cooling cell,hull plating,IHD,uranium,inactive isotopes (1heat/sec),active isotopes(uranium pulses)]


    That's 10 possibilities per slot. There are 7 possible configurations of reactor and cooling options/heating options around.


    Even just enumerating this I see -far- too many possibilities to attack by brute force. Even smarter brute force would still run in to far too many possibilities.

  • Well just checking which combinations work isnt gonna work not even with a super computer. There like 8^54 combinations for a 6 chamber reactor. Thats about 5,85*10^48 diferent configurations. How big is that? Well you can only have 1,84*10^19 RAM on a 64 bit system (ok in reality it will be less but in theory it should be possible).

  • Nah, wasn't intending on brute forcing it, but writing an AI program to do it. Still not going to be the fastest thing in the world, but it'll get there. Not as if I'll say: "Try everything, and hope it works." Basically, I'm going to work on a simple brute force, and add in shortcut after shortcut, like the efficiency of using as many slots as possible, the realization that fitting the entire thing with uranium cells probably isn't a good idea, the importance of more cooling cells, and less heat dispensers, while not melting things. Basically a copy of me, but a far more compliant-to-do-absurd-and-menial-tasks version. Sure it might take some time, and that time might be better spent elsewhere, but hell why not? On a separate note, if that planner was done in java it would save me a lot of time to throw me some of that code as a starter.

  • There's only a few uranium setups (like 2 2-pair uranium or 4 uranium in a row etc.) so a program that optimises HD and cooling cell design would be epic through that will also take away the fun of puzzling :).


    Also if you want a even better design the program should also look at the heat distribution. Through this wont make much of a diference (one of my reactors got improved from 24,75 cycles to near 25 cycles) its still a improvement :)

  • There's only a few uranium setups (like 2 2-pair uranium or 4 uranium in a row etc.) so a program that optimises HD and cooling cell design would be epic through that will also take away the fun of puzzling :).


    Also if you want a even better design the program should also look at the heat distribution. Through this wont make much of a diference (one of my reactors got improved from 24,75 cycles to near 25 cycles) its still a improvement :)

    Hah, i see you upgraded it even more, nice one. But effective EU/t actually dropped, isn't that worse ?

  • It could be because it rounds numbers or something i dunno. But the diference between 24,94 and 25 is noticable. Its the diference between a crater and a intact reactor after 25 cycles. Not that anyone will actualy run 25 cycles in a row with instant refuel through. Takes almost 70 hours to run 25 cycles. Before this reactor blows up some update prolly fucked up your world.