• Welcome to the Speedsolving.com, home of the web's largest puzzle community!
    You are currently viewing our forum as a guest which gives you limited access to join discussions and access our other features.

    Registration is fast, simple and absolutely free so please, join our community of 40,000+ people from around the world today!

    If you are already a member, simply login to hide this message and begin participating in the community!

How to make a good 5x5 solver?

abunickabhi

Member
Joined
Jan 9, 2014
Messages
6,706
Location
Yo
WCA
2013GHOD01
YouTube
Visit Channel
Currently there are a lot of robots that can solve Rubiks cube fast and optimally. But the problem is still difficult for big cubes, as finding the optimal solution using brute force takes time. Also making a lego solver for big cubes is a mechanical challenge too. How to approach this problem if I only have a bit of idea of CS brute force algorithms, basics of servo motors and a bit about scrambling programs.
 

xyzzy

Member
Joined
Dec 24, 2015
Messages
2,878
You almost never want optimal solving for any puzzle with more than a billion or so states. What you really want is usually a reasonably short solution, as opposed to the absolute shortest possible. Just like you don't solve a cube by solving everything simultaneously, you don't ask the computer to do that either. You solve a part of it, then another part, and repeat until it's done.

With only two exceptions (IAssemble's solver and Ed Trice's solver), the low-move-count methods for solving big cubes are all adaptations of the Kociemba 2-phase algorithm to big cubes. If you really want to write your own solver, start with reimplementing Kociemba 2-phase and understanding some of the theory behind it. If you don't mind using someone else's solver, Chen Shuang has a 444 solver and a 555 solver and Daniel Walton has an arbitrary-size solver.

I have an abandonware solver for 333, 444, and all odd sizes 555 and above with the best known average move count (in OBTM) for odd big cubes 777 and up. (The decryption key is intentionally missing the last character; you can easily find it by brute force if you really want it.) This also has a Redi Cube optimal solver, an early iteration of my Loopover 4×4 optimal solver, various bits of code that were used for the 555 God's number bound calculations, among other nonsense. Hey, I'm not a professional programmer; don't call me out for bad software development practices.
 

abunickabhi

Member
Joined
Jan 9, 2014
Messages
6,706
Location
Yo
WCA
2013GHOD01
YouTube
Visit Channel
Thanks for the reply. I have been watching cubesolving forum for the last 3 years and know about Ed and Daniel's solver. I have also checked out the binary exes of their solvers on the github repo (https://cubesolvingprograms.freeforums.net/board/6/5x5x5-solving-programs).

I see that Ed is trying AI solver nowadays. It is fun to follow that progress.
I will check out Chen's repo.

Is there any other workaround other than Kociemba's 2-phase algorithm?
 

Nir1213

Member
Joined
Aug 31, 2020
Messages
1,465
YouTube
Visit Channel
Thanks for the reply. I have been watching cubesolving forum for the last 3 years and know about Ed and Daniel's solver. I have also checked out the binary exes of their solvers on the github repo (https://cubesolvingprograms.freeforums.net/board/6/5x5x5-solving-programs).

I see that Ed is trying AI solver nowadays. It is fun to follow that progress.
I will check out Chen's repo.

Is there any other workaround other than Kociemba's 2-phase algorithm?
i think he talked about a 1 phase alg but its big i guess
 
Top