Puzzle - RPL and others

[Dave Britten]

(04-23-2021 07:30 AM)EdS2 Wrote:  I'll be interested to see the execution times: this puzzle can be solved using a determined brain and only a few case-by-case analyses. By brute force there are quite a few combinations, so I wonder if there are ways to short cut the search and make big improvements.

I put together a not-particularly-elegant 71B version that finds a solution in about 55 seconds, so that should set the bar for execution time.

EDIT: Got it down to 219 bytes and about 36 seconds. There's still room to shave off a few bytes here and there.

Code:

0010 OPTION BASE 1 @ DESTROY D @ DIM D(9),R
0020 CALL NX(1,0,D(),R) @ DISP R
0030 SUB NX(N,S,D(),R)
0040 I=N-MOD(S,N) @ S=S+I
0050 IF I>9 THEN STOP 
0060 IF D(I) THEN GOTO 110
0070 D(I)=1
0080 IF N<9 THEN CALL NX(N+1,S*10,D(),R) ELSE R=S
0090 IF R THEN STOP 
0100 D(I)=0
0110 I=I+N @ S=S+N
0120 GOTO 50
0130 END SUB