Generate an algorithm to get from one state to another.

[atmosopher]

piggybacking on this thread, does anyone know how software like Cube Explorer adapts the two-phase algorithm to solving/trying to reach positions with pieces/stickers "ignored"?

 

[Bruce MacKenzie]

piggybacking on this thread, does anyone know how software like Cube Explorer adapts the two-phase algorithm to solving/trying to reach positions with pieces/stickers "ignored"?

I really don't know for sure but I'll hazard a guess.

The first stage of the Kociemba two phase solver takes an arbitrary cube state to an element of the subgroup <U,D,L2,R2,F2,B2>. In an optimal solver once an element of the subgroup is found it is tested to see if it is the solved cube. If not the program continues to search deeper for first stage solutions until it finds one that also solves the whole cube. To ignore some cubies would simply involve modifying the test one applies to the 1st stage solution to ignore the desired cubies.

 

[Ben Whitmore]

piggybacking on this thread, does anyone know how software like Cube Explorer adapts the two-phase algorithm to solving/trying to reach positions with pieces/stickers "ignored"?

Pretty sure that cube explorer only does optimal solves when pieces are ignored, without using any sort of multiphase algorithm

 

[N's-cvt]

So let's say I want to generate a de-orientation algorithm of corners for ZZ-Blah and my A-point is R U R' and my B-point is the Pi oll (R U2 R2 U' R2 U' R2 U2 R), using the mathematical way of solving this S=B* A', I think my setup is R U2 R2 U' R2 U' R2 U2 R+R U' R' is this the case I enter in cube explorer to generate the algorithm for de-oreintation from that Last F2L pair?

 

[N's-cvt]

I was messing around with Cube Explorer lately and have tried developing a method to generate algs for turning any F2L case into say a WV or SV case so I could use those LSLL methods efficiently, does someone know the best way to generate the transformation algs?

nvm I got it worked out, however for those who want to know the method it is rather simple. First you define what point A F2L and point B F2L is then you consider B the solved state which means you need to get A to a solved state. Because CE generates algs to a solved cube state you imagine F2L B on a clear cube then take into account what colors those are in place of point B, then put in place of point A. CE will think that the corner and edge need to be solved which will give you the algs for the transformation.