Undecidable problems I There quite a few undecidable problems For - - PowerPoint PPT Presentation

undecidable problems i
SMART_READER_LITE
LIVE PREVIEW

Undecidable problems I There quite a few undecidable problems For - - PowerPoint PPT Presentation

Undecidable problems I There quite a few undecidable problems For example, program verification is in general not solvable We will discuss an undecidable example called the halting problem November 17, 2020 1 / 11 A TM I A TM = { M ,


slide-1
SLIDE 1

Undecidable problems I

There quite a few undecidable problems For example, program verification is in general not solvable We will discuss an undecidable example called the “halting problem”

November 17, 2020 1 / 11

slide-2
SLIDE 2

ATM I

ATM = {M, w | M : a TM that accepts w} We will prove that ATM is undecidable However, ATM is Turing recognizable We can simply simulate M, w To be decidable we hope to avoid an infinite loop if at one point, know it cannot halt ⇒ reject Thus this problem is called the halting problem

November 17, 2020 2 / 11

slide-3
SLIDE 3

Diagonalization method I

We need a technique called “diagonalization method” for the proof It was developed by Cantor in 1873 to check if two infinite sets are equal Example: consider set of even integers versus set of {0, 1}∗ Both are infinite sets. Which one is larger ? Definition: two sets are equal if elements can be paired

November 17, 2020 3 / 11

slide-4
SLIDE 4

Definition 4.12 I

f is a one-to-one function if: f (a) = f (b) if a = b x y x y Left: a one-to-one function; right: not

November 17, 2020 4 / 11

slide-5
SLIDE 5

Definition 4.12 II

f : A → B onto if ∀b ∈ B, ∃a such that f (a) = b Example: f (a) = a2, where A = (−∞, ∞) and B = (−∞, ∞) This is not an onto function because for b = −1, there is no a such that f (a) = b

November 17, 2020 5 / 11

slide-6
SLIDE 6

Definition 4.12 III

However, if we change it to f (a) = a2, where A = (−∞, ∞) and B = [0, ∞) it becomes an onto function Definition: a function is called a correspondence if it is one-to-one and onto Example: f (a) = a3, where A = (−∞, ∞) and B = (−∞, ∞)

November 17, 2020 6 / 11

slide-7
SLIDE 7

Example 4.13 I

N = {1, 2, . . .} E = {2, 4, . . .} The two sets can be paired n f (n) = 2n 1 2 2 4 . . . . . . We consider N and E have the same size Definition: a set is countable if it is finite or same size as N

November 17, 2020 7 / 11

slide-8
SLIDE 8

Rational Numbers Countable I

Q = {m/n | m, n ∈ N} countable . . .

5 1 4 1 3 1 2 1 1 1

. . .

5 2 4 2 3 2 2 2 1 2

. . .

5 3 4 3 3 3 2 3 1 3

. . .

5 4 4 4 3 4 2 4 1 4

. . .

5 5 4 5 3 5 2 5 1 5

. . .

5 6 4 6 3 6 2 6 1 6

· · · · · · · · · · · · · · ·

November 17, 2020 8 / 11

slide-9
SLIDE 9

Rational Numbers Countable II

(Latex source from https://divisbyzero.com/2013/04/16/ countability-of-the-rationals-drawn-using-tikz/)

Note that we skip counting elements with common factors (e.g., 2/2)

November 17, 2020 9 / 11

slide-10
SLIDE 10

Real Numbers not Countable I

We will use the diagonalization method The proof is by contradiction Assume R is countable. Then there is a table as follows n f (n) 1 3.14159 . . . 2 55.55555. . . 3 0.12345 . . . 4 0.50000 . . . . . .

November 17, 2020 10 / 11

slide-11
SLIDE 11

Real Numbers not Countable II

Consider x = 0.4641 . . . 4 = 1, 6 = 5 We have x = f (n), ∀n But x ∈ R, so a contradiction To avoid the problem 1 = 0.9999 · · · for every digit of x we should not choose 0 or 9

November 17, 2020 11 / 11