CS101 Data Structures
AVL Trees
Outline
Define height balancing
Maintaining balance within a tree
– AVL trees
– Difference of heights
– Maintaining balance after insertions and erases
– Can we store AVL trees as arrays?
Background
From previous lectures:
– Binary search trees store linearly ordered data
– Best case height: (ln(n))
– Worst case height: O(n)
Requirement:
– Define and maintain balance to ensure (ln(n)) height
Prototypical Examples
These two examples demonstrate how we can correct for
imbalances: starting with this tree, add 1:
Prototypical Examples
This is more like a linked list; however, we can fix this…
Prototypical Examples
Promote 2 to the root, demote 3 to be 2’s right child, and 1 remains
the left child of 2
Prototypical Examples
The result is a perfect, though trivial tree
Prototypical Examples
Alternatively, given this tree, insert 2
Prototypical Examples
Again, the product is a linked list; however, we can fix this, too
Prototypical Examples
Promote 2 to the root, and assign 1 and 3 to be its children
Prototypical Examples
The result is, again, a perfect tree
These examples may seem trivial, but they are the basis for the
corrections in the next data structure we will see: AVL trees
AVL Trees
Named after Adelson-Velskii and Landis
A binary search tree is said to be AVL balanced if:
– The difference in the heights between the left and right sub-trees is at
most 1, and
– Both sub-trees are themselves AVL trees
Recall:
– An empty tree has height –1
– A tree with a single node has height 0
AVL Trees
AVL trees with 1, 2, 3, and 4 nodes:
AVL Trees
Here is a larger AVL tree (42 nodes):
AVL Trees
The root node is AVL-balanced:
– Both sub-trees are of height 4:
AVL Trees
All other nodes (e.g., AF and BL) are AVL balanced
– The sub-trees differ in height by at most one
Height of an AVL Tree
By the definition of complete trees, any complete binary search tree
is an AVL tree
Thus, the upper bound on the number of nodes in an AVL tree of
height h is a perfect binary tree with 2h + 1 – 1 nodes
What is the lower bound?
Height of an AVL Tree
Let F(h) be the fewest number of nodes in a tree of height h
From a previous slide:
F(0) = 1
F(1) = 2
F(2) = 4
Can we find F(h)?
Height of an AVL Tree
The worst-case AVL tree of height h would have:
– A worst-case AVL tree of height h – 1 on one side,
– A worst-case AVL tree of height h – 2 on the other, and
– The root node
We get: F(h) = F(h – 1) + 1 + F(h – 2)
Height of an AVL Tree
This is a recurrence relation:
1 h0
F( h ) 2 h 1
F( h 1) F( h 2) 1 h 1
The solution?
– Note that F(h) + 1 = (F(h – 1) + 1) + (F(h – 2) + 1)
– Therefore, F(h) + 1 is a Fibonacci number:
F(0) + 1 = 2 → F(0) = 1
F(1) + 1 = 3 → F(1) = 2
F(2) + 1 = 5 → F(2) = 4
F(3) + 1 = 8 → F(3) = 7
F(4) + 1 = 13 → F(4) = 12
F(5) + 1 = 21 → F(5) = 20
F(6) + 1 = 34 → F(6) = 33
Height of an AVL Tree
Alternatively, if it wasn’t so simple:
> rsolve( {F(0) = 1, F(1) = 2,
F(h) = 1 + F(h - 1) + F(h - 2)}, F(h) );
h h
2 2
h h 2 5 2 5
3 5 1 1
5 1 3 5 1 1 5 1 5
5 1
10 2 2 2 2 10 2 2 5 ( 1 5 ) 5 ( 1 5 )
> asympt( %, h );
h (h I) h
( 3 5 5 ) ( 1 5 ) 1e ( 53 5 ) 2
1
h 5 h
5 ( 1 5 ) 2 ( 1 5 ) ( 1 5 )
h
1 5
c
2
Height of an AVL Tree
This is approximately
F(h) ≈ 1.8944 h – 1
where ≈ 1.6180 is the golden ratio
– That is, F(h) = (h)
Thus, we may find the maximum value of h for a given n:
n 1
log log n 1 1.3277 1.4404 lg n 1 1.3277
1.8944
Height of an AVL Tree
In this example, n = 88, the worst- and best-case scenarios differ in
height by only 2
Height of an AVL Tree
If n = 106, the bounds on h are:
– The minimum height: log2( 106 ) – 1 ≈ 19
– The maximum height: log( 106 / 1.8944 ) < 28
Maintaining Balance
Observe that:
– Inserting a node can increase the height of a tree by at most 1
– Removing a node can decrease the height of a tree by at most 1
This may cause some nodes to be unbalanced. We may need to
rebalance the tree after insertion or removal.
Maintaining Balance
To calculate changes in height, the member function height() must
run in (1) time
– Our implementation of height is (n)
– Introduce a member variable：
int tree_height;
– This variable is updated during inserting and erasing
Maintaining Balance
Only insert and erase may change the height
– This is the only place we need to update the height
– These algorithms are already recursive
Insert
template <typename Type>
bool AVL_node<Type>::insert( const Type & obj, AVL_node<Type> *&ptr_to_this ) {
if ( empty() ) {
ptr_to_this = new AVL_node( obj );
return true;
} else if ( obj < element ) {
if ( left()->insert( obj, left_tree ) ) {
tree_height = max( height(), 1 + left()->height() );
return true;
} else {
return false;
}
} else if ( obj > element ) {
// ...
} else {
return false;
}
}
Maintaining Balance
Consider this AVL tree
Maintaining Balance
Consider inserting 15 into this tree
– In this case, none of the heights of the trees change
Maintaining Balance
The tree remains balanced
Maintaining Balance
Consider inserting 42 into this tree
Maintaining Balance
Consider inserting 42 into this tree
– Now we see the heights of two sub-trees have increased by one
– The tree is still balanced
Maintaining Balance
If a tree is AVL balanced, for an insertion to cause an imbalance:
– The heights of the sub-trees must differ by 1
– The insertion must increase the height of the deeper sub-tree by 1
Maintaining Balance
Suppose we insert 23 into our initial tree
Maintaining Balance
The heights of each of the sub-trees from here to the root are
increased by one
Maintaining Balance
However, only two of the nodes are unbalanced: 17 and 36
Maintaining Balance
However, only two of the nodes are unbalanced: 17 and 36
Maintaining Balance
We can promote 23 to where 17 is, and make 17 the left child of 23
Maintaining Balance
Thus, that node is no longer unbalanced
– Incidentally, the root is now also balanced!
Maintaining Balance
Consider adding 6:
Maintaining Balance
The height of each of the trees in the path back to the root are
increased by one
Maintaining Balance
The height of each of the trees in the path back to the root are
increased by one
– However, only the root node is now unbalanced
Maintaining Balance
To fix this, we will look at the general case…
Maintaining Balance: Case 1
Consider the following setup
– Each blue triangle represents a tree of height h
Maintaining Balance: Case 1
Insert a into this tree: it falls into the left subtree BL of b
left-left
– Assume BL remains balanced
– The tree rooted at b is also balanced
Maintaining Balance: Case 1
The tree rooted at node f is now unbalanced
– We will correct the imbalance at this node
Maintaining Balance: Case 1
Here are examples of when the
insertion of 7 may cause this
situation when h = –1, 0, and 1
Maintaining Balance: Case 1
We will modify these three pointers
– At this point, this references the unbalanced root node f
AVL_node<Type> *b = left(),
*BR = b->right();
Maintaining Balance: Case 1
Specifically, we will rotate these two nodes around the root:
– Recall the first prototypical example
– Promote node b to the root and demote node f to be the right child of b
AVL_node<Type> *b = left(),
*BR = b->right();
Maintaining Balance: Case 1
This requires the address of node f to be assigned to the right_tree
member variable of node b
AVL_node<Type> *b = left(),
*BR = b->right();
b->right_tree = this;
Maintaining Balance: Case 1
Assign any former parent of node f to the address of node b
Assign the address of the tree BR to left_tree of node f
AVL_node<Type> *b = left(),
*BR = b->right();
b->right_tree = this;
ptr_to_this = b;
left_tree = BR;
Maintaining Balance: Case 1
The nodes b and f are now balanced, and all remaining nodes of the
subtrees are in their correct positions
Maintaining Balance: Case 1
Q: will this insertion affect the balance of any ancestors all the way
back to the root?
No, because height of the tree rooted at b equals the original height
of the tree rooted at f
Maintaining Balance: Case 1
In our example case, the correction
Maintaining Balance: Case 1
In our three sample cases
with h = –1, 0, and 1, the
node is now balanced
and the same height
as the tree before the
insertion
Maintaining Balance: Case 2
Alternatively, consider the insertion of c where b < c < f into our
original tree
Maintaining Balance: Case 2
Assume that the insertion of c increases the height of BR
left-right
– Once again, f becomes unbalanced
Maintaining Balance: Case 2
Here are examples of when the
insertion of 14 may cause this
situation when h = –1, 0, and 1
Maintaining Balance: Case 2
Unfortunately, the previous correction does not fix the imbalance at
the root of this sub-tree: the new root, b, remains unbalanced
Maintaining Balance: Case 2
In our three sample cases
with h = –1, 0, and 1,
doing the same thing
as before results in
a tree that is still
unbalanced…
– The imbalance is just
shifted to the other
side
Maintaining Balance: Case 2
We need to look into BR
Maintaining Balance: Case 2
Re-label BR as a tree rooted at d with two subtrees of height h – 1
Maintaining Balance: Case 2
Now an insertion causes an imbalance at f
– The addition of either c or e will cause this
Maintaining Balance: Case 2
We will reassign the following pointers
AVL_node<Type> *b = left(),
*d = b->right(),
*DL = d->left(),
*DR = d->right();
Maintaining Balance: Case 2
Specifically, we will order these three nodes as a perfect tree
– Recall the second prototypical example
AVL_node<Type> *b = left(),
*d = b->right(),
*DL = d->left(),
*DR = d->right();
Maintaining Balance: Case 2
To achieve this, b and f will be assigned as children of the new root
d
AVL_node<Type> *b = left(),
*d = b->right(),
*DL = d->left(),
*DR = d->right();
d->left_tree = b;
d->right_tree = this;
Maintaining Balance: Case 2
We also have to connect the two subtrees and original parent of f
AVL_node<Type> *b = left(),
*d = b->right(),
*DL = d->left(),
*DR = d->right();
d->left_tree = b;
d->right_tree = this;
ptr_to_this = d;
b->right_tree = DL;
left_tree = DR;
Maintaining Balance: Case 2
Now the tree rooted at d is balanced
Maintaining Balance: Case 2
Again, the height of the root did not change
Maintaining Balance: Case 2
In our three sample cases
with h = –1, 0, and 1, the
node is now balanced
and the same height
as the tree before the
insertion
Maintaining balance: symmetric cases
There are two symmetric cases to those we have examined:
– Insertions into the right-right sub-tree
– Insertions into either the right-left sub-tree
Insert (Implementation)
template <typename Type>
void AVL_node<Type>::insert( const Type & obj, AVL_node<Type> *ptr_to_this ) {
if ( empty() ) {
ptr_to_this = new AVL_node<Type>( obj );
return true;
} else if ( obj < element ) {
if ( left() -> insert() ) {
if ( left()->height() - right()->height() == 2 ) {
// determine if it is a left-left or left-right insertion
// perform the appropriate correction
} else {
tree_height = std::max( height(), left()->height() );
}
return true;
} else {
return false;
}
} else if ( obj > element ) {
// ...
Insertion (Implementation)
Comments:
– Both balances are (1)
– All insertions are still (ln(n))
– It is possible to tighten the previous code
– Aside
• if you want to explicitly rotate the nodes A and B, you must also pass a
reference to the parent pointer as an argument:
insert( Type obj, AVL_node<Type> * & parent )
Insertion
Consider this AVL tree
Insertion
Insert 73
Insertion
The node 81 is unbalanced
– A left-left imbalance
Insertion
The node 81 is unbalanced
– A left-left imbalance
Insertion
The node 81 is unbalanced
– A left-left imbalance
– Promote the intermediate node to the imbalanced node
Insertion
The node 81 is unbalanced
– A left-left imbalance
– Promote the intermediate node to the imbalanced node
– 75 is that node
Insertion
The node 81 is unbalanced
– A left-left imbalance
– Promote the intermediate node to the imbalanced node
– 75 is that node
Insertion
The tree is AVL balanced
Insertion
Insert 77
Insertion
The node 87 is unbalanced
– A left-right imbalance
Insertion
The node 87 is unbalanced
– A left-right imbalance
Insertion
The node 87 is unbalanced
– A left-right imbalance
– Promote the intermediate node to the imbalanced node
Insertion
The node 87 is unbalanced
– A left-right imbalance
– Promote the intermediate node to the imbalanced node
– 81 is that value
Insertion
The node 87 is unbalanced
– A left-right imbalance
– Promote the intermediate node to the imbalanced node
– 81 is that value
Insertion
The tree is balanced
Insertion
Insert 76
Insertion
The node 78 is unbalanced
– A left-left imbalance
Insertion
The node 78 is unbalanced
– Promote 77
Insertion
Again, balanced
Insertion
Insert 80
Insertion
The node 69 is unbalanced
– A right-left imbalance
– Promote the intermediate node to the imbalanced node
Insertion
The node 69 is unbalanced
– A right-left imbalance
– Promote the intermediate node to the unbalanced node
– 75 is that value
Insertion
Again, balanced
Insertion
Insert 74
Insertion
The node 72 is unbalanced
– A right-right imbalance
– Promote the intermediate node to the imbalanced node
– 75 is that value
Insertion
The node 72 is unbalanced
– A right-right imbalance
– Promote the intermediate node to the imbalanced node
Insertion
Again, balanced
Insertion
Insert 64?
Insertion
This causes no imbalance
Insertion
Insert 55
Insertion
The node 69 is imbalanced
– A left-left imbalance
– Promote the intermediate node to the imbalanced node
Insertion
The node 69 is imbalanced
– A left-left imbalance
– Promote the intermediate node to the imbalanced node
– 63 is that value
Insertion
The tree is now balanced
Insertion
Insert 70?
Insertion
The root node is now imbalanced
– A right-left imbalance
– Promote the intermediate node to the root
– 75 is that value
Insertion
The root node is imbalanced
– A right-left imbalance
– Promote the intermediate node to the root
– 63 is that node
Insertion
The result is AVL balanced
Erase
Removing a node from an AVL tree may cause more than one AVL
imbalance
– Like insert, erase must check if it caused an imbalance
– Unfortunately, it may cause O(h) imbalances that must be corrected
• Insertions will only cause one imbalance that must be fixed
Erase
Consider the following AVL tree
Erase
Suppose we erase the front node: 1
Erase
While its previous parent, 2, is not unbalanced, its grandparent 3 is
– The imbalance is in the right-right subtree
Erase
We can correct this with a simple balance
Erase
The node of that subtree, 5, is now balanced
Erase
Recursing to the root, however, 8 is also unbalanced
– This is a right-left imbalance
Erase
Promoting 11 to the root corrects the imbalance
Erase
At this point, the node 11 is balanced
Erase
Still, the root node is unbalanced
– This is a right-right imbalance
Erase
Again, a simple balance fixes the imbalance
Erase
The resulting tree is now AVL balanced
– Note, few erases will require one balance, even fewer will require more
than one
Time complexity
Insertion
– May require one correction to maintain balance
– Each correction require (1) time
Erasion
– May require O(h) corrections to maintain balance
– Each correction require (1) time
– Depth h is ( ln(n) )
– So the time complexity is O( ln(n) )
AVL Trees as Arrays?
We previously saw that:
– Complete tree can be stored using an array using (n) memory
– An arbitrary tree of n nodes requires O(2n) memory
Is it possible to store an AVL tree as an array and not require
exponentially more memory?
AVL Trees as Arrays?
Recall that in the worst case, an AVL tree of n nodes has a height at
most
log(n) – 1.3277
Such a tree requires an array of size
2log(n) – 1.3277 + 1 – 1
We can rewrite this as
2–0.3277 nlog(2) ≈ 0.7968 n1.44
Thus, we would require O(n1.44) memory
AVL Trees as Arrays?
While the polynomial behaviour of n1.44 is not as bad as exponential
behaviour, it is still reasonably sub-optimal when compared to the
linear growth associated with link-
allocated trees
Here we see n and n1.44 on [0, 1000]
Summary
AVL trees
– Balance: difference in subtree heights is 0 or 1
– Insertion and erasion may require the tree to be rebalanced by tree
rotations
Usage Notes
• These slides are made publicly available on the web for anyone to
use
• If you choose to use them, or a part thereof, for a course at another
institution, I ask only three things:
– that you inform me that you are using the slides,
– that you acknowledge my work, and
– that you alert me of any mistakes which I made or changes which you
make, and allow me the option of incorporating such changes (with an
acknowledgment) in my set of slides
Sincerely,
Douglas Wilhelm Harder, MMath
dwharder@alumni.uwaterloo.ca
