Switching Circuits and Logic Design (CS21202)

Module 6
Binary Decision Diagrams

Prof. Indranil Sengupta

Dr. Monosij Maitra

Department of Computer Science and Engineering

IIT Kharagpur
Binary Decision Diagrams (BDD)

Switching Circuits and Logic Design (CS20202) 2

Binary Decision Diagrams (BDD)

• A data structure used to represent a Boolean function.

• Represented as a rooted, directed, acyclic graph, which consists of
several decision nodes and two terminal nodes, called 0-terminal and
1-terminal.
• Each decision node is labeled by a Boolean variable and has two child nodes
called low child, and high child.
• The edge from a node to a low (high) child represents an assignment of the
variable to 0 (1).
• A BDD is said to be ordered if different variables appear in the same
order on all paths from the root. 3
 Example a b c f
f 0 0 0 1
0 0 1 0
a 0 1 0 0
0 1 1 1
b b 1 0 0 0
1 0 1 0
1 1 0 1
c c c c 1 1 1 1

1 0 0 1 0 0 1 1

4
• Construction of a BDD is based on the Shannon expansion of a
function.

5
 Shannon Expansion
• Given a Boolean function f (x1, x2, …, xi , …, xn)
• Positive cofactor
fi1 = f (x1, x2, …, 1, …, xn)
• Negative cofactor
fi0 = f (x1, x2, …, 0, …, xn)
• Shannon’s expansion theorem states that
f = xi fi0 + xi fi1
f = (xi + fi0 )(xi + fi1 )

6
How to construct BDD?

f = ac + bc + a’b’c’
= a’ (b’c’ + bc) + a (c + bc) f This is the first step.

= a’ (b’c’ + bc) + a (c) The process is continued

a for all input variables.

Low High
child child
b’c’ + bc c

7
 f = ac + bc + abc
= a (bc + bc) + a (c + bc) Expand by a
= a (bc + bc) + a (c)

b(c) + b(c) b(c) + b(c) Expand by b

Expand by c
c(1) + c(0) c(0) + c(1) c(0) + c(1) c(0) + c(1)

Variable ordering: a, b, c

8
 f

b b

c c c c

1 0 0 1 0 1 0 1

9
Another Example: f = a’bc + b’c’ + ac’

10
Another Example: f = a’bc + b’c’ + ac’
f = a’bc + b’c’ + ac’ Expand by c
= c .(a + b’) + c.(a’b)

11
Another Example: f = a’bc + b’c’ + ac’
f = a’bc + b’c’ + ac’ Expand by c
= c .(a + b’) + c.(a’b)

a.(b) + a.(1) a(b) + a(0) Expand by a

12
 Another Example: f = a’bc + b’c’ + ac’
f = a’bc + b’c’ + ac’ Expand by c
= c .(a + b’) + c.(a’b)

a.(b) + a.(1) a(b) + a(0) Expand by a

b(1) + b(0) b(1) + b(1) b(0) + b(1) b(0) + b(0) Expand by b

13
 Another Example: f = a’bc + b’c’ + ac’
f = a’bc + b’c’ + ac’ Expand by c
= c .(a + b’) + c.(a’b)

a.(b) + a.(1) a(b) + a(0) Expand by a

b(1) + b(0) b(1) + b(1) b(0) + b(1) b(0) + b(0) Expand by b

Variable ordering: c, a, b

14
Another Example: f = a’bc + b’c’ + ac’
f

a a

b b b b

1 0 1 1 0 1 0 0
15
Another Example: f = xy + yz + zx

16
 Another Example: f = xy + yz + zx
f = xy + yz + zx Expand by x
= x .(yz) + x.(1.y + 1.z + y.z)

y.(0) + y.(z) y(z) + y(1) Expand by y

z(0) + z(0) z(0) + z(1) z(0) + z(1) z(1) + z(1) Expand by z

Variable ordering: x, y, z

17
Another Example: f = xy + yz + zx
f

y y

z z z z

0 0 0 1 0 1 1 1
18
Another Example: f = w1’ w3’ + w1w2 + w1w3

19
 Another Example: f = w1’ w3’ + w1w2 + w1w3
f = w1’w3’ + w1w2 + w1w3 Expand by w1
= w1 .(0.w2 + 1.w3‘) + w1.(1. w2 + 1. w3)
= w1 .(w3‘) + w1.(w2 + w3)

w2‘.(w3‘) + w2.(w3‘) w2 (w3) + w2(1) Expand by w2

w3(1) + w3(0) w3(1) + w3(0) w3(0) + w3(1) w3(1) + w3(1) Expand by w3

Variable ordering: w1, w2, w3 20

Another Example: f = w1’ w3’ + w1w2 + w1w3
f

w1

w2 w2

w3 w3 w3 w3

1 0 1 0 0 1 1 1
21
 Variable Ordering (OBDD)
• The size of a BDD is determined both by the function being represented and the
chosen ordering of the variables.
• For some functions, the size of a BDD may vary between a linear to an exponential range
depending upon the ordering of the variables.
• An example:
f(x1,…,x2n) = x1x2 + x3x4 + … + x2n-1x2n

Variable ordering: x1 < x3 < … < x2n-1 < x2 < x4 < … < x2n
BDD requires 2n+1 nodes to represent the function.

Variable ordering: x1 < x2 < x3 < x4 < … < x2n-1 < x2n
BDD requires 2n nodes to represent the function.

22
BDD for the function f(x1, ...,
x8) = x1x2 + x3x4 + x5x6 + x7x8
using bad variable ordering

23
Same function using good variable
ordering

24
• Important point to note:
• It is essential to find a good variable ordering when using the OBDD data
structure in practice.
• The problem of finding the best variable ordering is NP-hard.
• Several heuristics for variable ordering have been proposed.

25
Reduced Ordered BDD (ROBDD)

• An ordered binary decision diagram is said to be reduced ordered BDD (ROBDD) if

the following two graph reduction rules are applied:
• Merge any isomorphic subgraphs.
• Eliminate any node whose two children are isomorphic.

26
Example: f = w1’ w3’ + w1w2 + w1w3
f

w1

w2 w2

w3 w3 w3 w3

1 0 1 0 0 1 1 1
27
Example: f = w1’ w3’ + w1w2 + w1w3
f

w1

w2 w2

w3 Merge two w3 w3 w3
subgraphs

1 0 1 0 0 1 1 1
28
Example: f = w1’ w3’ + w1w2 + w1w3
f
Eliminate w1
node w2

w2 w2

w3 Merge two w3 w3 w3
subgraphs

1 0 1 0 0 1 1 1
29
Reduced Ordered BDD (ROBDD)

• An ordered binary decision diagram is said to be reduced ordered BDD (ROBDD) if

the following two graph reduction rules are applied:
• Merge any isomorphic subgraphs.
• Eliminate any node whose two children are isomorphic.

• The advantage of an ROBDD is that it is canonical (unique) for a given function.

• There is exactly one ROBDD for a given variable ordering.
• This property makes it useful in checking functional equivalence.

30
Some Properties of ROBDD

• The following properties hold in a ROBDD:

a) Uniqueness: No two distinct nodes u and v are labeled with the same
variable name and have the same low and high successor.
b) Non-redundant: No variable node u has identical low and high successor.

31
Reduction Rule: Merge Isomorphic Subtrees

x x x

y z y z y z

x x x

y z y z y z

32
Reduction Rule: Remove Redundant Nodes

y y

33
Construction of ROBDD: an example
Given function: f (x1, x2, x3) = x1.x2 + x1’.x3 + x1.x2’.x3

34
 x1 x1
0 1 0 1

x2 x2 x2 x2
0 1 0 1 0 1 0 1
x3 x3 x3 x3 x3 x3 x3
x3 1 0 0
0 1 0 1 0 1 0 1 0 1
1 1
0 1 0 1 0 1 1 1 0
0 1

x1 1 x1
0 1
x2 x2 0
1 0 x2
0
0
x3 1 1 x3 1 1
0 0
0 1 0 1

35
Some Benefits of BDD

• Checking for tautology is trivial.

• BDD is a constant 1.
• Complementation:
• Given a BDD for a function f, the BDD for f can be obtained by simply
interchanging the terminal nodes.
• Equivalence check:
• Two functions f and g are equivalent if their BDDs (under the same variable
ordering) are the same.

36
Use of BDD in Synthesis
• BDD is canonical for a given variable ordering.
• It implicitly uses factored representation:

x
x’h + xh = h
h y h y

a b a b

h x h x h x
ah + bh = (a+b)h
y z y z y z
• Variable reordering can reduce the size of BDD.
• Implicit logic minimization.

• Some redundancy is also removed during the construction of BDD itself.

38
MUX realization of functions

f f

x x
0 1

f g
f g

39
MUX-based Functional Decomposition
f f

0 1
h

g
h
f g f

An example ===>

40
 f f

a ab
0 1
b a
c d
c b
d
0 1
0 1

41
A Complete Mapping Example

x1
1
0
0 x2

x3 1 1
0

0 1

42
To Summarize

• BDDs have been used traditionally to represent and manipulate

Boolean functions.
• Used in synthesis systems.
• Used in formal verification tools.
• Efficient packages to manipulate BDDs are available.

43