Recursive Definitions and

Structural Induction

Niloufar Shafiei
Recursion
Sometimes it is difficult to define an object
explicitly.

It may be easy to define this object in terms of

itself.

This process is called recursion.

1
Recursion
We can use recursion to define sequences,
functions, and sets.

Example:
an=2n for n = 0,1,2,…
1,2,4,8,16,32,…
After giving the first term, each term of the sequence
can be defined from the previous term.
a1=1 an+1 = 2an

2
Recursion
When a sequence is defined recursively,
mathematical induction can be used to prove
results about the sequence.

Let P(k) be proposition about ak.

Basis step:

Verify P(1).
Inductive step:

Show k
1 (P(k)
P(k+1)).

3
Recursively defined functions
Assume f is a function with the set of nonnegative
integers as its domain
We use two steps to define f.
Basis step:

Specify the value of f(0).
Recursive step:

Give a rule for f(x) using f(y) where 0
y<x.

Such a definition is called a recursive or inductive

definition.
4
Example
Suppose

f(0) = 3

f(n+1) = 2f(n)+2,  n
0.
Find f(1), f(2) and f(3).

Solution:
f(1) =
2f(0) + 2 = 2(3) + 2 = 8
f(2) =
2f(1) + 2 = 2(8) + 2 = 18
f(3) =
2f(2) + 2 = 2(18) + 2 = 38
5
Example
Give an inductive definition of the factorial function F(n) = n!.
Solution:

Basis step: (Find F(0).)
F(0)=1

Recursive step: (Find a recursive formula for F(n+1).)
F(n+1) = (n+1) F(n)

What is the value of F(5)?
F(5) = 5F(4)
= 5 . 4F(3)
= 5 . 4 . 3F(2)
= 5 . 4 . 3 . 2F(1)
= 5 . 4 . 3 . 2 . 1F(0)
= 5 . 4 . 3 . 2 . 1 . 1 = 120
6
Recursive functions

Recursively defined functions should be well

defined.

It means for every positive integer, the value

of the function at this integer is determined
in an unambiguous way.

7
Example
Assume a is a nonzero real number and n is a
nonnegative integer.
Give a recursive definition of an.
Solution:

Basis step: (Find F(0).)
F(0) = a0 = 1

Recursive step: (Find a recursive formula for
F(n+1).)
F(n+1) = a . an = a . F(n)
8
Example
n
Give a recursive definition of
ak.
K=0
Solution:

Basis step: (Find F(0).)
0
F(0) =
ak = a0
K=0

Recursive step: (Find a recursive formula for F(n+1).)

F(n+1) = F(n) + an+1

9
Recursive functions
In some recursive functions,

The values of the function at the first k positive integers
are specified

A rule is given to determine the value of the function at
larger integer from its values at some of the preceding k
integers.

Example:
f(0) = 2 and f(1) = 3
f(n+2) = 2f(n) + f(n+1) + 5,  n
0.
10
Fibonacci numbers

The Fibonacci numbers, f0, f1, f2, …, are

defined by the equations

f0 = 0

f1 = 1

fn = fn-1 + fn-2
for n = 2,3,4,… .

11
Example
Find the Fibonacci number f4.

Solution:
f4 = f3 + f2
f2 = f0 + f1 = 0 + 1 = 1
f3 = f1 + f2 = 1 + 1 = 2

f4 = f3 + f2 = 1 + 2 = 3

12
Example
Show that whenever n
3, fn >
n-2, where
=(1 +

5)/2. (Hint:
2=
+1)
Proof by strong induction:

Find P(n)
P(n) is fn >
n-2.

Basis step: (Verify P(3) and P(4) are true.)
f3 >
1
2 > (1 +
5) / 2
f4 >
2
3 > (1 +
5)2 / 4
13
 Example
Show that whenever n3, fn > n-2, where =(1 +
5)/2. (Hint: 2= +1)
Proof by strong induction:
 Inductive step: (Show k ([P(3)P(4)…P(k)] P(k+1)) is true.)

Inductive hypothesis:
fj > j-2 when 3
j
k.

Show  k
4 P(k+1) is true. (Show fk+1 > k-1 is true.)
 Let k
4.
fk+1 = fk + fk-1
By induction hypothesis, fk > k-2 and fk-1 > k-3.
fk+1 = fk + fk-1 > k-2 + k-3 =  . k-3 + k-3 = (+1) . k-3
= 2 . k-3 = k-1
We showed P(k+1) is true, so by strong induction fn > n-2 is true.

14
Recursively defined sets and
structures
Assume S is a set.
We use two steps to define the elements of S.
Basis step:

Specify an initial collection of elements.
Recursive step:

Give a rule for forming new elements
from those already known to be in S.

15
Example
Consider S  Z defined by
Basis step: (Specify initial elements.)
3
S
Recursive step: (Give a rule using existing elements)
If x
S and y
S, then x+y
S.
3
S
3+3=6
S
6+3=9
S
6 + 6 = 12
S
…
16
Example
Show that the set S defined in previous slide, is the set of all
positive integers that are multiples of 3.
Solution:
 Let A be the set of all positive integers divisible by 3.
 We want to show that A=S
 Part 1: (Show AS using mathematical induction.)

Show x (x  A
x  S).

Define P(n).
P(n) is “3n  S”.

Basis step: (Show P(1).)
P(1) is “3  S”.
By recursive definition of S, 3  S, so P(1) is true.
17
Example
Show that the set S defined in previous slide, is the set of all positive
integers that are multiples of 3.
Solution:
 Part 1: (Show AS using mathematical induction.)

Inductive step: (Show k
1 P(k)P(k+1).)
 Define inductive hypothesis:
P(k) is “3k  S”.
 Show k
1 P(k+1) is true.
P(k+1) is “3(k+1)  S”.
3(k+1) = 3k + 3
By recursive definition of S, since 3  S and 3k  S, (3k+3) S.
By mathematical induction, n
1 3n  S.

18
 Example
Show that the set S defined in previous slide, is the set of all positive
integers that are multiples of 3.
Solution:
 Part 2: (Show SA.)

Show x (x  S
x  A)
 By basis step, 3  S.
Since 3 is positive multiple of by 3, 3  A.
 By recursive step, If x  S and y  S, then x+y  S.

Show If x  A and y  A, then x+y  A.

Assume x  A and y  A, so x and y are positive
multiples of 3.

So, x+y is positive multiple of 3 and x+y  A.
So, S=A.

19
Set of strings
Finite sequences of form a1,a2,…,an are called strings.
The set * of strings over the alphabet  can be defined by
Basic step: 
*
( is the empty string containing no symbols.)
Recursive step:
If w
* and x
, then wx
.

Example:
={0,1}

*
0 = 0
* 1 = 1
*
01
* 11
*
010
* 110
*
20
Concatenation
Let  be a set of symbols and * be a set of strings formed from symbols
in .
The concatenation of two strings, denoted by ., recursively as follows.
Basic step:
If w
*, then w. = w.
Recursive step:
If w1
* and w2
* and x
, then w1 .(w2x) = (w1 . w2 ) x.

Example:
w1 = abc w2 = def
w1 . w2 = w1w2 = abcdef

21
Example
Give a recursive definition of l(w), the length of the string w.
Solution:
Basis step:
l() = 0
Recursive step:
l(wx) =
l(w) + 1, where w
* and x
.

Example:
l(ab) = l(a) + 1
= l() + 1 + 1
=0+1+1=2
22
Structural induction

Instead of mathematical induction to prove a

result about a recursively defined sets, we
can used more convenient form of
induction known as structural induction.

23
Structural induction
Assume we have recursive definition for the set S.
Let n  S.
Show P(n) is true using structural induction:
Basis step:

Assume j is an element specified in the basis step of the
definition.

Show j P(j) is true.
Recursive step:

Let x be a new element constructed in the recursive step of
the definition.

Assume k1, k2, …, km are elements used to construct an
element x in the recursive step of the definition.

Show  k1, k2, …, km ((P(k1) P(k2)… P(km))
P(x)).
24
Example
Use structural induction, to prove that l(xy) = l(x)+l(y),
where x
* and y
*.
Proof by structural induction:
 Define P(n).
P(n) is l(xn) = l(x)+l(n) whenever x
*.
 Basis step: (P(j) is true, if j is specified in basis step of the
definition.)

Show P() is true.

P() is l(x) = l() + l(x).

Since x = x, l(x) = l(x)
= l(x) + 0 = l(x) + l()

So, P() is true.
25
Example
Use structural induction, to prove that l(xy) = l(x)+l(y), where x
 * and y  *.
Proof by structural induction:
 Inductive step: (P(y)
P(ya) where a)

Inductive hypothesis:(P(y))
 l(xy) = l(x) + l(y)

Show that P(ya) is true.
 Show l(xya) = l(x) + l(ya)
 By recursive definition, l(xya) = l(xy) + 1.
 By inductive hypothesis, l(xya)= l(x) + l(y) + 1.
 By recursive definition (l(ya)= l(y) + 1), l(xya)= l(x) + l(ya).
 So, P(ya) is true.

26
Example
Well-formed formulae for compound propositions:
Basis step: T(true), F(false) and p, where p is a propositional
variable, are well-formed.
Recursive step: If F and E are well-formed formulae, then (¬
E), (EF), (EF), (E
F) and (EF) are well-formed
formulae.

Example:
p  F is well-formed.
p ¬
q is not well-formed.

27
Example
Well-formed formulae for operations:
Basis step: x, where x is a numeral or variable, is well-formed.
Recursive step: If F and E are well-formed formulae, then
(E+F), (E-F), (E*F), (E/F) and (E
F) are well-formed
formulae.
(* denotes multiplication and
denotes exponentiation.)

Example:
3 + (5-x) is well-formed.
3 * + x is not well-formed.

28
Example
Show that well-formed formulae for compound propositions
contains an equal number of left and right parentheses.
Proof by structural induction:
 Define P(x)
P(x) is “well-formed compound proposition x contains an
equal number of left and right parentheses”
 Basis step: (P(j) is true, if j is specified in basis step of the
definition.)

T, F and propositional variable p is constructed in the
basis step of the definition.

Since they do not have any parentheses, P(T), P(F)
and P(p) are true.

29
Example
Proof by structural induction:
 Recursive step:

Assume p and q are well-formed formulae.

Let lp be the number of left parentheses in p.

Let rp be the number of right parentheses in p.

Let lq be the number of left parentheses in q.

Let rq be the number of right parentheses in q.

Assume lp= rp and lq= rq.

We need to show that (¬p), (pq), (pq), (p
q) and (p
q) also contains an equal number of left and right
parentheses.

30
Example
Proof by structural induction:

Recursive step:

The number of left parentheses in (¬p) is lp+1 and
the number of right parentheses in (¬p) is rp+1.

Since lp= rp, lp+1= rp+1 and (¬p) contains an equal
number of left and right parentheses.

The number of left parentheses in other compund
propositions is lp+ lq+ 1 and the number of right
parentheses in (¬p) is rp+ rq+ 1.

Since lp= rp and lq= rq, lp+ lq+ 1 = rp+ rq+ 1 and other
compound propositions contain an equal number of
left and right parentheses.
So, by structural induction, the statement is true.
31
Structural induction

Structural induction is really just a version of

(strong) induction.

32
Tree
A graph is made up vertices and edges connecting some pairs
of vertices.

A tree is a special type of a graph.

A rooted tree consists of a set of vertices a distinguished
vertex called root and edges connecting these vertices. (A
tree has no cycle.)
root

33
Rooted tree
The set of rooted trees can be defined recursively
by these steps:
Basis step:
A single vertex r is a rooted tree.
Recursive step:
Suppose that T1, T2, …, Tn are disjoint rooted
trees with roots r1, r2, …, rn, respectively.
Then, the graph formed by starting with a root r
which is not in any of the rooted tree T1, T2, …, Tn,
and adding an edge from r to each of the vertices
r1, r2, …, rn, is also a rooted tree.
34
Rooted tree
r

root root

T1 T2

35
Rooted tree
Basis step:

Inductive step:

36
Extended binary trees
The set of extended binary trees can be defined
recursively by these steps:
Basis step:
The empty set is an extended binary tree.
Recursive step:
Assume T1 and T2 are disjoint extended binary
trees.
Then, there is an extended binary tree, denoted
T1 . T2, consisting of a root r together with edges
connecting the roots of left subtree T1 and the
right subtree T2 when these trees are nonempty.
37
Extended binary trees
The root of an extended binary tree is
connected to at most two subtrees.
Basis step:

Inductive step:

38
Full binary trees
The set of full binary trees can be defined
recursively by these steps:
Basis step:
There is a full binary tree consisting only of a
single vertex r.
Recursive step:
Assume T1 and T2 are disjoint full binary trees.
Then, there is a full binary tree, denoted T1 . T2,
consisting of a root r together with edges
connecting the root to each of the roots of the left
subtree T1 and the right subtree T2.
39
Full binary tree
The root of a full binary tree is connected to
exactly two subtrees.
Basis step:
Inductive step:

40
Height of full binary trees
We define height h(T) of a full binary tree T recursively.

Basis step:

Assume T is a full binary tree consisting of a single
vertex.

h(T)=0
Recursive step:

Assume T1 and T2 are full binary trees.

h(T1 . T2) = 1 + max (h(T1),h(T2))

41
Height of full binary trees
h(T) = 0

h(T) = 1+ max(0,0) = 1

h(T) = 1+ max(1,1) = 2

42
Number of vertices of full binary
trees
We define number of vertices n(T) of a full binary tree T
recursively.

Basis step:

Assume T is a full binary tree consisting of a single
vertex.

n(T)=1
Recursive step:

Assume T1 and T2 are full binary trees.

n(T1 . T2) = 1 + n(T1) + n(T2)
43
Number of vertices of full binary
trees
n(T) = 1

n(T) = 1+ 1 + 1 = 3

n(T) = 1+ 3 + 3 = 7

44
Structural induction
How to show a result about full binary trees using
structural induction?

Basis step:
Show that the result is true for the tree consisting
of a single vertex.
Recursive step:
Show that if the result is true for trees T1 and T2,
then it is true for T1 . T2, consisting of a root r
which has T1 as its left subtree and T2 as its right
subtree.
45
Example
Show if T is a full binary tree T, then n(T)
2h(T)+1 - 1.
Proof by structural induction:
 Basis step:

Assume T is a full binary tree consisting of a single
vertex.

Show n(T)
2h(T)+1 - 1 is true.
1
20+1 - 1=1

So, it is true for T.
 Inductive step:

Assume T1 and T2 are full binary trees.

Assume n(T1)
2h(T1)+1 - 1 and n(T2)
2h(T2)+1 - 1 are true.

Assume T= T1 . T2.
46
Example
Show if T is a full binary tree T, then n(T)
2h(T)+1 - 1.
Proof by structural induction:
 Inductive step:

Show n(T)
2h(T)+1 - 1 is true.

By recursive definition, n(T) = 1 + n(T1) + n(T2).

By recursive definition, h(T) = 1 + max(h(T1),h(T2)).
n(T) = 1 + n(T1) + n(T2)

1 + 2h(T1)+1 - 1 + 2h(T2)+1 - 1 (by inductive hypothesis)

2 . max(2h(T1)+1,2h(T2)+1) - 1
= 2 . 2max ( h(T1),h(T2) ) + 1 - 1 (max(2x,2y) = 2max(x,y))
= 2 . 2h(T) - 1 (by recursive definition)
= 2h(T)+1 - 1
47
Recommended exercises
1,3,5,7,9,21,23,25,27,29,33,35,44,57,59

48