I have a very high dimensional transition table built for a Turing machine which generates alphabet strings as outputs. Now, I'm trying to make another Turing machine that decodes the output and finds the original input given to the machine.
Is there any efficient way to trace down the rules to find the input? The order in which the rules were applied initially is very important and trying out every possible rule was not ideal. I've thought of doing dynamic programming but this still takes a huge amount of time.