cyb0124 / fissionopt

Nuclearcraft fission reactor optimizer

Web version: https://leu-235.com/ (Built using emscripten. See web/compile.bat for details.)

To build the benchmark, clone xtl and xtensor to the parent directory of this repo and run CMake.

fissionopt's Issues

Different values when import json to NC Reactor planner

Seems like the values for FuelCells differ from when i load my own Nuclearcraft.cfg, is there any way to change the Cooler\Cell attributes in this generator ?

Active cooler is not accessible

Somehow an active cooler was placed in non-accessible position with no way to pipe fluids to it from inside.

S1 Size: 7x7x7
S2 Fuel: HECm-243 @ Default config
S4 Rates configured according to DD&SS modpack config where passive cooler Qz = 90 and all active coolers has 2x default rates
S5 Default parameters

I run generator with this config and left (forgot) for some time, stopping it on episode 819.
Full page capture

Both website and compiled js not working

I'm not sure what is causing the error message I'm getting. I'm on Arch Linux. I tried the normal website, I compiled it myself with the required libs using Emscripten. I also tried both versions with Firefox and Chromium. Both of them give roughly the browser console errors, the following is via Firefox:

*each newline is a new message unless tabbed

GEThttps://leu-235.com/main.css
GEThttps://leu-235.com/FissionOpt.js
GEThttps://leu-235.com/main.js
no native wasm support detected FissionOpt.js:9:4100
jQuery.Deferred exception: WebAssembly is not defined FissionOpt@https://leu-235.com/FissionOpt.js:9:4165
@https://leu-235.com/main.js:1:11
e@https://code.jquery.com/jquery-3.5.0.slim.min.js:2:30211
Deferred/then/l/</t<@https://code.jquery.com/jquery-3.5.0.slim.min.js:2:30513
setTimeout handler*Deferred/then/l/<@https://code.jquery.com/jquery-3.5.0.slim.min.js:2:30722
c@https://code.jquery.com/jquery-3.5.0.slim.min.js:2:28500
fireWith@https://code.jquery.com/jquery-3.5.0.slim.min.js:2:29245
fire@https://code.jquery.com/jquery-3.5.0.slim.min.js:2:29281
c@https://code.jquery.com/jquery-3.5.0.slim.min.js:2:28500
fireWith@https://code.jquery.com/jquery-3.5.0.slim.min.js:2:29245
ready@https://code.jquery.com/jquery-3.5.0.slim.min.js:2:32218
F@https://code.jquery.com/jquery-3.5.0.slim.min.js:2:31997
EventListener.handleEvent*@https://code.jquery.com/jquery-3.5.0.slim.min.js:2:32366
@https://code.jquery.com/jquery-3.5.0.slim.min.js:2:220
@https://code.jquery.com/jquery-3.5.0.slim.min.js:2:225
 undefined jquery-3.5.0.slim.min.js:2:31766
Uncaught ReferenceError: WebAssembly is not defined
    FissionOpt https://leu-235.com/FissionOpt.js:9
    <anonymous> https://leu-235.com/main.js:1
    jQuery 13
FissionOpt.js:9:4165
    FissionOpt https://leu-235.com/FissionOpt.js:9
    <anonymous> https://leu-235.com/main.js:1
    jQuery 13
GEThttps://leu-235.com/favicon.ico
[HTTP/1.1 200 OK 0ms]

When simply using the website normally with a mouse, the experience is that clicking any links/"buttons" changes nothing.

fluid transport inside reactors.

i am having some problems figuring out how to move fluids inside the reactor to the right places. you cant use fluid transport from active cooler to active cooler so do you move the fluids trough other active fluid coolers that have no use for cooling (not having the right requirements). and if you need to use active fluid coolers to transport it then if possible change the air gaps for active fluid coolers with maby a setting button.

Faster! FASTER! (And maybe more efficient)

Is there any way to make the website work faster? It barely takes up any CPU time for me, and I still think it's slow.

As for efficiency? I'm sure that there is a worked out "Perfect Reactor" for a given size, so what if there was a database of basic settings reactors. No custom values or limits on coolers, just the basic settings.
It won't be small, at 480 designs, or 3,840 if we take into account the symmetry setting, and double that if we allow active coolers, or fuck it, 251,658,240 with every single active cooler combination, but in my opinion, just the 960 non-symmetric, both with and without active cooling, is sufficient. Some extra optimizations will need to be taken, like having it discard rotation for same sized sides, if not done already, but in my ignorant opinion, it should be possible, no?

how do you know whe its done?

i have run the program but I dont know when it is doen cous it keep going with the episode, stages and iteration?

Any information on how much progress has been made?

My 5x7x5 generator had been running for hours, so I limited it to just a few cooler types but I have no idea how close it is to reaching a solution... How do I interpret the bar graph? Is there any way of knowing how many minutes / hours / decades the thing is going to run? What are some general expectations regarding compute time for various configurations? Any way of at least getting a suboptimal solution faster?

Thanks!

Suggestion: Weights for optimisation function

Firstly, thank you for this tool. The community has been desperately needing such a tool.

We are currently given three choices for optimisation:

1. Power
2. Breeder
3. Efficiency

Some of the designs generated by efficiency are nice, but I'd like slightly more power. It think it would be a nice user-feature to have some sort of weighting to the total value to be optimised. Something like 0.2power, 0.8efficiency would yield a less efficient reactor, but it may give just enough RF/t that I'm looking for.

When using RL, this is where a dynamic function has its advantages (if you're deriving new policy every run? I've not checked out your RL code). You can solve what is essentially an entirely different optimisation problem with an infinite space.

two active coolers of diffrent sorts toching.

ive made a 24x24x24 reactor and there are some active coolers touching other active coolers and that isnt possible.

you can see that the helium, redstone and emerald are touching cryothium.

.json export has borked fuel stats and can't be imported

example in the rar
reactor(1) is 5x3x5 with HEU-235
reactor(2) is 7x7x7 with HEN
both have no Fuel name as stats so ThizPlanner can't load them sadly
hellrage has no real problem with it just that fuel-stats are empty besides the rf and heat gain

tmp.zip

Compile as standalone executable

Can this be compiled into a standalone executable for faster execution for larger reactors?