Abstract model for an algorithm solving a problem, Theory of Computation

These assumptions hold for addition, for instance. Every instance of addition has a unique solution. Each instance is a pair of numbers and the possible solutions include any third number. We can systematically list all instances along with all possible solutions by systematically listing all triples of numbers. This is not completely trivial-we can't, for instance, list all triples starting with 0 and then all triples starting with 1, etc. Since there are in?nitely many triples starting with zero, we would never get around to listing any starting with one. Suppose, though, that we are only concerned with the Natural Numbers, {0, 1, . . .}. If we ?rst list all triples that sum to zero (i.e., just the triple h0, 0, 0i) and then all triples that sum to one (i.e., h1, 0, 0i, h0, 1, 0i, h0, 0, 1i), etc., we are guaranteed that we will eventually list any given triple.

With the exception of the assumption that the solution is unique (which can be fudged in a variety of ways) these assumptions are pretty nearly minimal. We can't even consider solving a problem algorithmically unless every instance has a solution. An algorithm must produce some answer for every instance. If there is no answer for some instance, then whatever answer it produces will necessarily be wrong. (Note that if we modify the problem to require that we return "No Solution" in the case that none exists, we will have converted it into a problem that has a solution for every instance-albeit one that sometimes has the solution "No Solution".) The third assumption is true of every reasonable problem. In fact, it takes a fairamount of the theory of computation to even get to the point where we can argue that problems that don't satisfy the assumption might exist. Under these assumptions we can reduce our model to a machine for checking the correctness of solutions:

1809_Abstract model for an algorithm solving a problem.png

Posted Date: 3/20/2013 5:56:13 AM | Location : United States







Related Discussions:- Abstract model for an algorithm solving a problem, Assignment Help, Ask Question on Abstract model for an algorithm solving a problem, Get Answer, Expert's Help, Abstract model for an algorithm solving a problem Discussions

Write discussion on Abstract model for an algorithm solving a problem
Your posts are moderated
Related Questions
write short notes on decidable and solvable problem

The Myhill-Nerode Theorem provided us with an algorithm for minimizing DFAs. Moreover, the DFA the algorithm produces is unique up to isomorphism: every minimal DFA that recognizes

. On July 1, 2010, Harris Co. issued 6,000 bonds at $1,000 each. The bonds paid interest semiannually at 5%. The bonds had a term of 20 years. At the time of issuance, the market r

Theorem (Myhill-Nerode) A language L ⊆ Σ is recognizable iff ≡L partitions Σ* into ?nitely many Nerode equivalence classes. Proof: For the "only if" direction (that every recogn

design a turing machine that accepts the language which consists of even number of zero''s and even number of one''s?


Kleene called this the Synthesis theorem because his (and your) proof gives an effective procedure for synthesizing an automaton that recognizes the language denoted by any given r

In Exercise 9 you showed that the recognition problem and universal recognition problem for SL2 are decidable. We can use the structure of Myhill graphs to show that other problems

design a tuning machine for penidrome