MATH 302 Lecture 15

From Notes
Jump to navigation Jump to search

« previous | Wednesday, October 19, 2011 | next »


Fibonacci Sequence

Recursive definition

fib1 n

   | n == 0 = 0
   | n == 1 = 1
   | otherwise = fib1 (n-1) + fib1 (n-2)

Easy to describe, but very expensive to evaluate:

fib1(6) |- fib1(5) | |- fib1(4) | | |- fib1(3) | | | |- fib1(2) | | | | |- fib1(1) | | | | `- fib1(0) | | | `- fib1(1) | | `- fib1(2) | | |- fib1(1) | | `- fib1(0) | `- fib1(3) | |- fib1(2) | | |- fib1(1) | | `- fib1(0) | `- fib1(1) `- fib1(4)

  |- fib1(3)
  |  |- fib1(2)
  |  |  |- fib1(1)
  |  |  `- fib1(0)
  |  `- fib1(1)
  `- fib1(2)
     |- fib1(1)
     `- fib1(0)

Iterative definition

def fib2 n

 return 0 if n == 0
 return 1 if n == 1
 x,y=0,1
 1.upto(n-1) { x,y = y,x+y }
 y

end

Counting Problems

Towers of Hanoi

Moving disks from one peg to another, etc. Example: It takes 7 moves to shift 3 disks to another peg.

Conjecture: It takes moves to move disks

Let represent the number of moves for disks

Moving disks takes . Move the last disk to another peg, then move the pegs on top of it:

Look at pattern of values for :

1 1
2 3
3 7
4 15

This supports the conjecture:

Proof by induction.


Basis step. true.

Inductive step. Assume that for some , . By the recurrence relation, we have

Proving the inductive step. The general result follows by the principle of mathematical induction.


Bit Strings

Count the number of bit strings of length with no consecutive zeros:

Possibilities
0 {} = 1
1 {1, 0} = 2
2 {11, 10, 01} = 3
3 {111, 110, 101, 011, 010} = 5
4 {1111, 1110, 1101, 1011, 0111, 1010, 0101, 0110} = 8

Constructing a Recurrence Relation

Given a string

For if is a 0, then must be an arbitrary legal n-1 string. If is a 1, then must be a legal n-1 string that ends in a 1. This means that must be a valid n-2 strings (conforming to ).

Count the number of possibilities for the last digit in the string. Get to a point where must be a legal string, then add all of the cases to form the case. If there is ever a case where must be an illegal string, then subtract the number of legal strings from the number of total possibilities