theory - How does a non deterministic turing machine work? -

I understand that they are not real, and whenever they have an option, even when there are 2 options, they Are calculated. But, for example, if I say:

"There is a non-deterministic assumption that the graph from the graph to the article is the vertebrae p" (the reference graph is isomorphism)

< P> What could this mean? I understand bisection, but it says "estimates without any estimated". If this is guessing, then how is it an algorithmic approach? How can this guarantee how it is going to work?

There are different ways of a picture one I have useful Oracle model Have you ever seen a side cartoon Have seen where there is a derivative on the blackboard "There is a miracle here" which is one of the intermediate steps? In this version of NDTM, when you have to choose something, Oracle writes the correct version on the right part of the tape. (This Grey End Johnson, Computers and Integrity , is derived from his classic book on NP-complete problems.) You are not allowed to believe that you got the right one, however, is not.

Therefore, when you think in a safe way, then you are getting the right scenario for your purposes, provided that someone exists.

This is not a good basis for an algorithm, because the complexity of implementing a non-deterministic Turing machine is basically very high in noneterminalistic states, and according to the algorithmic equivalent nodetemnostic estimation of every possible To launch is to try.

To look at the theoretical perspective, I will translate it as "such a bijection that ...." Get another chapter from another algorithm point, or the second chapter of the same book, because that approach is also useless for minor large graphs.