My solver-2x2

[dariodsa]

I am building my own solver of rubik's cube 2x2. I have made it, but I have problem with time. I have to wait 45 seconds to get right sequences of moves. So my question is which method can I use to speed it up?
I use brute force with some improvement.

 

[vcuber13]

the best method is to not use brute force

(3*6)^11=6.4*10^14 things to run through

 

[dariodsa]

How do you mean vcuber13?

 

[Stefan]

(3*6)^11=6.4*10^14 things to run through

That's not brute force. That's intentionally naive brute force. Doesn't take much smartness to at least do (3*3) * (2*3)^10 = 5.4*10^8 instead. That's what I would call a reasonable brute force method. And actually yours is 6.4*10^13.

 

[moralsh]

You could just try to implement one of the known methods, they won't be optimal but will be more optimal than brutw force!

 

[dariodsa]

Is there any list of sequence of all moves that could solve cube 2x2?

 

[vcuber13]

That's not brute force. That's intentionally naive brute force. Doesn't take much smartness to at least do (3*3) * (2*3)^10 = 5.4*10^8 instead. That's what I would call a reasonable brute force method. And actually yours is 6.4*10^13.

could you explain your number? my mistake it is to the 13.

You could just try to implement one of the known methods, they won't be optimal but will be more optimal than brutw force!

brute force should be optimal

 

[moralsh]

it could give an optimal solution, but any (computer) algorithm using brute force is not optimal

 

[Stefan]

could you explain your number?

Only turn R,U,F, never L/D/B, and never a side twice in a row. So for the first turn you have 3*3 possibilities and for each subsequent turn 2*3 possibilities.

 

[vcuber13]

Only turn R,U,F, never L/D/B, and never a side twice in a row. So for the first turn you have 3*3 possibilities and for each subsequent turn 2*3 possibilities.

i forgot that 2x2 only RUF moves are required

 

[Zarxrax]

It would be really great if your solver could treat opposite colors as one, to create algorithms for guimond-like methods. This would give it a unique quality over other 2x2 solvers that exists.

 

[Meep]

It would be really great if your solver could treat opposite colors as one, to create algorithms for guimond-like methods. This would give it a unique quality over other 2x2 solvers that exists.

But we can already do that by blanking stickers on a 3x3 solver.

 

[cuBerBruce]

It would be really great if your solver could treat opposite colors as one, to create algorithms for guimond-like methods. This would give it a unique quality over other 2x2 solvers that exists.

But we can already do that by blanking stickers on a 3x3 solver.

Being able to blank stickers does not give you the capability of simulating a tri-color cube.

EDIT: OK, Guimond method steps don't really require tri-color ability. Guimond orientation step only requires that you can define corner orientation, and Guimond separation only requires specifying white/yellow corner stickers. Interestingly, Cube Explorer complains if you only specify white/yellow corner stickers.