Boolean Algebra CHAPTER 15

Transcription

1 CHAPTER 15 Boolean Algebra 15.1 INTRODUCTION Both sets and propositions satisfy similar laws, which are listed in Tables 1-1 and 4-1 (in Chapters 1 and 4, respectively). These laws are used to define an abstract mathematical structure called a Boolean algebra, which is named after the mathematician George Boole ( ) BASIC DEFINITIONS Let B be a nonempty set with two binary operations + and, a unary operation, and two distinct elements 0 and 1. Then B is called a Boolean algebra if the following axioms hold where a, b, c are any elements in B: [B 1 ] Commutative laws: (1a) a + b = b + a (1b) a b = b a [B 2 ] Distributive laws: (2a) a + (b c) = (a + b) (a + c) (2b) a (b + c) = (a b) + (a c) [B 3 ] Identity laws: (3a) a + 0 = a (3b) a 1 = a [B 4 ] Complement laws: (4a) a + a = 1 (4b) a a = 0 We will sometimes designate a Boolean algebra by B,+,,, 0, 1 when we want to emphasize its six parts. We say 0 is the zero element, l, is the unit element, and a is the complement of a. We will usually drop the symbol and use juxtaposition instead. Then (2b) is written a(b + c) = ab + ac which is the familiar algebraic identity of rings and fields. However, (2a) becomes a + bc = (a + b)(a + c), which is certainly not a usual identity in algebra. The operations +,, and are called sum, product, and complement, respectively. We adopt the usual convention that, unless we are guided by parentheses, has precedence over, and has precedence over +. For example, a + b c means a + (b c) and not (a + b) c; a b means a (b ) and not (a b) Of course when a + b c is written a + bc then the meaning is clear. EXAMPLE 15.1 (a) Let B ={0, 1}, the set of bits (binary digits), with the binary operations of + and and the unary operation defined by Fig Then B is a Boolean algebra. (Note simply changes the bit, i.e., 1 = 0 and 0 = 1.) Copyright 2007, 1997, 1976 by The McGraw-Hill Companies, Inc. Click here for terms of use. 368

2 CHAP. 15] BOOLEAN ALGEBRA 369 Fig (b) Let B n = B B B (n factors) where the operations of +,, and are defined componentwise using Fig For notational convenience, we write the elements of B n as n-bit sequences without commas, e.g., x = and y = belong to B n. Hence x + y = , x y = , x = Then B n is a Boolean algebra. Here 0 = is the zero element, and 1 = is the unit element. We note that B n has 2 n elements. (c) Let D 70 ={1, 2, 5, 7, 10, 14, 35, 70}, the divisors of 70. Define +,, and on D 70 by a + b = lcm(a, b), a b = gcd(a, b), a = 70 a Then D 70 is a Boolean algebra with 1 the zero element and 70 the unit element. (d) Let C be a collection of sets closed under the set operations of union, intersection, and complement. Then C is a Boolean algebra with the empty set as the zero element and the universal set U as the unit element. Subalgebras, Isomorphic Boolean Algebras Suppose C is a nonempty subset of a Boolean algebra B. We say C is a subalgebra of B if C itself is a Boolean algebra (with respect to the operations of B). We note that C is a subalgebra of B if and only if C is closed under the three operations of B, i.e., +,, and. For example, {1, 2, 35, 70} is a subalgebra of D 70 in Example 15.1(c). Two Boolean algebras B and B are said to be isomorphic if there is a one-to-one correspondence f: B B which preserves the three operations, i.e., such that, for any elements, a, b in B, f(a+ b) = f(a)+ f(b), f(a b) = f(a) f(b) and f(a ) = f(a) 15.3 DUALITY The dual of any statement in a Boolean algebra B is the statement obtained by interchanging the operations + and, and interchanging their identity elements 0 and 1 in the original statement. For example, the dual of (1 + a) (b + 0) = b is (0 a) + (b 1) = b Observe the symmetry in the axioms of a Boolean algebra B. That is, the dual of the set of axioms of B is the same as the original set of axioms. Accordingly, the important principle of duality holds in B. Namely, Theorem 15.1 (Principle of Duality): The dual of any theorem in a Boolean algebra is also a theorem. In other words, if any statement is a consequence of the axioms of a Boolean algebra, then the dual is also a consequence of those axioms since the dual statement can be proven by using the dual of each step of the proof of the original statement.

3 370 BOOLEAN ALGEBRA [CHAP BASIC THEOREMS Using the axioms [B 1 ] through [B 4 ], we prove (Problem 15.5) the following theorem. Theorem 15.2: Let a, b, c be any elements in a Boolean algebra B. (i) (ii) (iii) (iv) Idempotent laws: (5a) a + a = a (5b) a a = a Boundedness laws: (6a) a + 1 = 1 (6b) a 0 = 0 Absorption laws: (7a) a + (a b) = a (7b) a (a + b) = a Associative laws: (8a) (a + b) + c = a + (b + c) (8b) (a b) c = a (b c) Theorem 15.2 and our axioms still do not contain all the properties of sets listed in Table 1-1. The next two theorems give us the remaining properties. Theorem 15.3: Let a be any element of a Boolean algebra B. (i) (Uniqueness of Complement) If a + x = 1 and a x = 0, then x = a. (ii) (Involution law) (a ) = a. (iii) (9a) 0 = 1. (9b) 1 = 0. Theorem 15.4 (DeMorgan s laws): (10a) (a + b) = a b. (10b) (a b) = a + b. We prove these theorems in Problems 15.6 and BOOLEAN ALGEBRAS AS LATTICES By Theorem 15.2 and axiom [B 1 ], every Boolean algebra B satisfies the associative, commutative, and absorption laws and hence is a lattice where + and are the join and meet operations, respectively. With respect to this lattice, a + 1 = 1 implies a 1 and a 0 = 0 implies 0 a, for any element a B. Thus B is a bounded lattice. Furthermore, axioms [B 2 ] and [B 4 ] show that B is also distributive and complemented. Conversely, every bounded, distributive, and complemented lattice L satisfies the axioms [B 1 ] through [B 4 ]. Accordingly, we have the following Alternate Definition: A Boolean algebra B is a bounded, distributive and complemented lattice. Since a Boolean algebra B is a lattice, it has a natural partial ordering (and so its diagram can be drawn). Recall (Chapter 14) that we define a b when the equivalent conditions a + b = b and a b = a hold. Since we are in a Boolean algebra, we can actually say much more. Theorem 15.5: The following are equivalent in a Boolean algebra: (1) a + b = b, (2) a b = a, (3) a + b = 1, (4) a b = 0 Thus in a Boolean alegbra we can write a b whenever any of the above four conditions is known to be true. EXAMPLE 15.2 (a) Consider a Boolean algebra of sets. Then set A precedes set B if A is a subset of B. Theorem 15.4 states that if A B then the following conditions hold: (1) A B = B (2)A B = A (3)A c B = U (4) A B c =

4 CHAP. 15] BOOLEAN ALGEBRA 371 (b) Consider the Boolean algebra D 70. Then a precedes b if a divides b. In such a case, lcm(a, b) = b and gcd(a, b) = a. For example, let a = 2 and b = 14. Then the following conditions hold: (1) lcm(2, 14) = 14. (3) lcm(2, 14) = lcm(35, 14) = 70. (2) gcd(2, 14) = 2. (4) gcd(2, 14 ) = gcd(2, 5) = REPRESENTATION THEOREM Let B be a finite Boolean algebra. Recall (Section 14.10) that an element a in B is an atom if a immediately succeeds 0, that is if 0 a. Let A be the set of atoms of B and let P (A) be the Boolean algebra of all subsets of the set A of atoms. By Theorem 14.8, each x = 0inB can be expressed uniquely (except for order) as the sum ( join) of atoms, i.e., elements of A. Say, x = a 1 + a 2 + +a r is such a representation. Consider the function f: B P (A) defined by f(x)={a 1,a 2,...,a r } The mapping is well defined since the representation is unique. Theorem 15.6: The above mapping f: B P (A) is an isomorphism. Thus we see the intimate relationship between set theory and abstract Boolean algebras in the sense that every finite Boolean algebra is structurally the same as a Boolean algebra of sets. If a set A has n elements, then its power set P (A) has 2 n elements. Thus the above theorem gives us our next result. Corollary 15.7: A finite Boolean algebra has 2 n elements for some positive integer n. EXAMPLE 15.3 Consider the Boolean algebra D 70 ={1, 2, 5,...,70} whose diagram is given in Fig. 15-2(a). Note that A ={2, 5, 7} is the set of atoms of D 70. The following is the unique representation of each nonatom by atoms: 10 = 2 5, 14 = 2 7, 35 = 5 7, 70 = Figure 15-2(b) gives the diagram of the Boolean algebra of the power set P (A) of the set A of atoms. Observe that the two diagrams are structurally the same. Fig SUM-OF-PRODUCTS FORM FOR SETS This section motivates the concept of the sum-of-products form in Boolean algebra by an example of set theory. Consider the Venn diagram in Fig of three sets A, B, and C. Observe that these sets partition the

5 372 BOOLEAN ALGEBRA [CHAP. 15 Fig rectangle (universal set) into eight numbered sets which can be represented as follows: (1) A B C (3) A B c C (5) A B c C c (7) A c B c C (2) A B C c (4) A c B C (6) A c B C c (8) A c B c C c Each of these eight sets is of the form A B C where: A = A or A c, B = B or B c, C = C or C c Consider any nonempty set expression E involving the sets A, B, and C, say, E =[(A B c ) c (A c C c )] [(B c C) c (A C c )] Then E will represent some area in Fig and hence will uniquely equal the union of one or more of the eight sets. Suppose we now interpret a union as a sum and an intersection as a product. Then the above eight sets are products, and the unique representation of E will be a sum (union) of products. This unique representation of E is the same as the complete sum-of-products expansion in Boolean algebras which we discuss below SUM-OF-PRODUCTS FORM FOR BOOLEAN ALGEBRAS Consider a set of variables (or letters or symbols), say x 1,x 2,...,x n.aboolean expression E in these variables, sometimes written E(x 1,...,x n ), is any variable or any expression built up from the variables using the Boolean operations +,, and. (Naturally, the expression E must be well-formed, that is, where + and are used as binary operations, and is used as a unary operation.) For example, E 1 = (x + y z) + (xyz + x y) and E 2 = ((xy z + y) + x z) are Boolean expressions in x, y, and z. A literal is a variable or complemented variable, such as x, x, y, y, and so on. A fundamental product is a literal or a product of two or more literals in which no two literals involve the same variable. Thus xz, xy z, x, y, x yz are fundamental products, but xyx z and xyzy are not. Note that any product of literals can be reduced to either 0 or a fundamental product, e.g., xyx z = 0 since xx = 0 (complement law), and xyzy = xyz since yy = y (idempotent law). A fundamental product P 1 is said to be contained in (or included in) another fundamental product P 2 if the literals of P 1 are also literals of P 2. For example, x z is contained in x yz, but x z is not contained in xy z since x is not a literal of xy z. Observe that if P 1 is contained in P 2, say P 2 = P 1 Q, then, by the absorption law, Thus, for instance, x z + x yz = x z. P 1 + P 2 = P 1 + P 1 Q = P 1

6 CHAP. 15] BOOLEAN ALGEBRA 373 Definition 15.1: A Boolean expression E is called a sum-of-products expression if E is a fundamental product or the sum of two or more fundamental products none of which is contained in another. Definition 15.2: Let E be any Boolean expression. A sum-of-products form of E is an equivalent Boolean sum-of-products expression. EXAMPLE 15.4 Consider the expressions E 1 = xz + y z + xyz and E 2 = xz + x yz + xy z Although the first expression E 1 is a sum of products, it is not a sum-of-products expression. Specifically, the product xz is contained in the product xyz. However, by the absorption law, E 1 can be expressed as E 1 = xz + y z + xyz = xz + xyz + y z = xz + y z This yields a sum-of-products form for E 1. The second expression E 2 is already a sum-of-products expression. Algorithm for Finding Sum-of-Products Forms Figure 15-4 gives a four-step algorithm which uses the Boolean algebra laws to transform any Boolean expression into an equivalent sum-of-products expression. Fig EXAMPLE 15.5 Suppose Algorithm 15.1 is applied to the following Boolean expression: E = ((xy) z) ((x + z)(y + z )) Step 1. Using DeMorgan s laws and involution, we obtain E = (xy + z )((x + z) + (y + z ) ) = (xy + z )(xz + yz) E now consists only of sums and products of literals. Step 2. Using the distributive laws, we obtain E = xyxz + xyyz + xz z + yzz E now is a sum of products. Step 3. Using the commutative, idempotent, and complement laws, we obtain E = xyz + xyz + xz + 0 Each term in E is a fundamental product or 0.

7 374 BOOLEAN ALGEBRA [CHAP. 15 Step 4. The product xz is contained in xyz ; hence, by the absorption law, xz + (xz y) = xz Thus we may delete xyz from the sum. Also, by the identity law for 0, we may delete 0 from the sum. Accordingly, E = xyz + xz E is now represented by a sum-of-products expression. Complete Sum-of-Products Forms A Boolean expression E = E(x 1,x 2,...,x n ) is said to be a complete sum-of-products expression if E is a sum-of-products expression where each product P involves all the n variables. Such a fundamental product P which involves all the variables is called a minterm, and there is a maximum of 2 n such products for n variables. The following theorem applies. Theorem 15.8: Every nonzero Boolean expression E = E(x 1,x 2,...,x n ) is equivalent to a complete sum-of-products expression and such a representation is unique. The above unique representation of E is called the complete sum-of-products form of E. Algorithm 15-1 in Fig tells us how to transform E into a sum-of-products form. Figure 15-5 contains an algorithm which transforms a sum-of-products form into a complete sum-of-products form. Fig EXAMPLE 15.6 Express E(x,y,z) = x(y z) into its complete sum-of-products form. (a) Apply Algorithm 15.1 to E so E is represented by a sum-of-products expression: (b) Now apply Algorithm 15.2 to obtain: E = x(y z) = x(y + z ) = xy + xz E = xy(z + z ) + xz (y + y ) = xyz + xyz + xyz + xy z = xyz + xyz + xy z Now E is reprsented by its complete sum-of-products form. Warning: The terminology in this section has not been standardized. The sum-of-products form for a Boolean expression E is also called the disjunctive normal form or DNF of E. The complete sum-of-products form for E is also called the full disjunctive normal form, or the disjunctive canonical form, or the minterm canonical form of E.

8 CHAP. 15] BOOLEAN ALGEBRA MINIMAL BOOLEAN EXPRESSIONS, PRIME IMPLICANTS There are many ways of representing the same Boolean expression E. Here we define and investigate a minimal sum-of-products form for E. We must also define and investigate prime implicants of E since the minimal sum-of-products involves such prime implicants. Other minimal forms exist, but their investigation lies beyond the scope of this text. Minimal Sum-of-Products Consider a Boolean sum-of-products expression E. Let E L denote the number of literals in E (counted according to multiplicity), and let E S denote the number of summands in E. For instance, suppose E = xyz + x y t + xy z t + x yzt Then E L = = 14 and E S = 4. Suppose E and F are equivalent Boolean sum-of-products expressions. We say E is simpler than F if: (i) E L <F L and E S F L, or (ii) E L F L and E S <F L We say E is minimal if there is no equivalent sum-of-products expression which is simpler than E. We note that there can be more than one equivalent minimal sum-of-products expressions. Prime Implicants A fundamental product P is called a prime implicant of a Boolean expression E if P + E = E but no other fundamental product contained in P has this property. For instance, suppose One can show (Problem 15.15) that: E = xy + xyz + x yz xz + E = E but x + E = E and z + E = E Thus xz is a prime implicant of E. The following theorem applies. Theorem 15.9: A minimal sum-of-products form for a Boolean expression E is a sum of prime implicants of E. The following subsections give a method for finding the prime implicants of E based on the notion of the consensus of fundamental products. This method can then be used to find a minimal sum-of-products form for E. Section gives a geometric method for finding such prime implicants. Consensus of Fundamental Products Let P 1 and P 2 be fundamental products such that exactly one variable, say x k, appears uncomplemented in one of P 1 and P 2 and complemented in the other. Then the consensus of P 1 and P 2 is the product (without repetitions) of the literals of P 1 and the literals of P 2 after x k and x k are deleted. (We do not define the consensus of P 1 = x and P 2 = x.) The following lemma (proved in Problem 15.19) applies. Lemma 15.10: Suppose Q is the consensus of P 1 and P 2. Then P 1 + P 2 + Q = P 1 + P 2. EXAMPLE 15.7 Find the consensus Q of P 1 and P 2 where: (a) P 1 = xyz s and P 2 = xy t. Delete y and y and then multiply the literals of P 1 and P 2 (without repetition) to obtain Q = xz st.

9 376 BOOLEAN ALGEBRA [CHAP. 15 (b) P 1 = xy and P 2 = y. Deleting y and y yields Q = x. (c) P 1 = x yz and P 2 = x yt. No variable appears uncomplemented in one of the products and complemented in the other. Hence P 1 and P 2 have no consensus. (d) P 1 = x yz and P 2 = xyz. Each of x and z appear complemented in one of the products and uncomplemented in the other. Hence P 1 and P 2 have no consensus. Consensus Method for Finding Prime Implicants Figure 15-6 contains an algorithm, called the consensus method, which is used to find the prime implicants of a Boolean expression E. The following theorem gives the basic property of this algorithm. Theorem 15.11: The consensus method will eventually stop, and then E will be the sum of its prime implicants. Fig EXAMPLE 15.8 Let E = xyz + x z + xyz + x y z + x yz. Then: E = xyz + x z + xyz + x y z (x yz includes x z ) = xyz + x y + xyz + x y z + xy (consensus of xyz and xyz ) = x z + x y z + xy (xyz and xyz include xy) = x z + x y z + xy + x y (consensus of x z and x y z) = x z + xy + x y (x y z includes x y ) = x z + xy + x y + yz (consensus of x z and xy) Now neither step in the consensus method will change E. Thus E is the sum of its prime implicants, which appear in the last line, that is, x z, xy, x y, and yz. Finding a Minimal Sum-of-Products Form The consensus method (Algorithm 15.3) is used to express a Boolean expression E as a sum of all its prime implicants. Figure 15-7 contains an algorithm which uses such a sum to find a minimal sum-of-products form for E.

10 CHAP. 15] BOOLEAN ALGEBRA 377 Fig EXAMPLE 15.9 We apply Algorithm 15.4 to the following expression E which (by Example 15.8) is now expressed as the sum of all its prime implicants: E = x z + xy + x y + yz Step 1. Express each prime implicant of E as a complete sum-of-products to obtain: x z = x z (y + y ) = x yz + x y z xy = xy(z + z ) = xyz + xyz x y = x y (z + z ) = x y z + x y z yz = yz (x + x ) = xyz + x yz Step 2. The summands of x z are x yz and x y z which appear among the other summands. Thus delete x z to obtain E = xy + x y + yz The summands of no other prime implicant appear among the summands of the remaining prime implicants, and hence this is a minimal sum-of-products form for E. In other words, none of the remaining prime implicants is superfluous, that is, none can be deleted without changing E LOGIC GATES AND CIRCUITS Logic circuits (also called logic networks) are structures which are built up from certain elementary circuits called logic gates. Each logic circuit may be viewed as a machine L which contains one or more input devices and exactly one output device. Each input device in L sends a signal, specifically, a bit (binary digit), 0 or 1 to the circuit L, and L processes the set of bits to yield an output bit. Accordingly, an n-bit sequence may be assigned to each input device, and L processes the input sequences one bit at a time to produce an n-bit output sequence. First we define the logic gates, and then we investigate the logic circuits. Logic Gates There are three basic logic gates which are described below. We adopt the convention that the lines entering the gate symbol from the left are input lines and the single line on the right is the output line. (a) OR Gate: Figure 15-8(a) shows an OR gate with inputs A and B and output Y = A + B where addition is defined by the truth table in Fig. 15-8(b). Thus the output Y = 0 only when inputs A = 0 and B = 0. Such an OR gate may, have more than two inputs. Figure 15-8(c) shows an OR gate with four inputs, A, B, C, D, and output Y = A + B + C + D. The output Y = 0 if and only if all the inputs are 0.

11 378 BOOLEAN ALGEBRA [CHAP. 15 Fig Suppose, for instance, the input data for the OR gate in Fig (c) are the following 8-bit sequences: A = , B = , C = , The OR gate only yields 0 when all input bits are 0. This occurs only in the 2nd, 5th, and 7th positions (reading from left to right). Thus the output is the sequence Y = (b) AND Gate: Figure 15-9(a) shows an AND gate with inputs A and B and output Y = A B (or simply Y = AB) where multiplication is defined by the truth table in Fig. 15-9(b). Thus the output Y = 1 when inputs A = 1 and B = 1; otherwise Y = 0. Such anand gate may have more than two inputs. Figure 15-9(c) shows an AND gate with four inputs, A, B, C, D, and output Y = A B C D. The output Y = 1ifand only if all the inputs are 1. Suppose, for instance, the input data for the AND gate in Fig. 15-9(c) are the following 8-bit sequences: A = , B = , C = , D = The AND gate only yields 1 when all input bits are 1. This occurs only in the 2nd, 3rd, and 7th positions. Thus the output is the sequence Y = Fig (c) NOT Gate: Figure 15-10(a) shows a NOT gate, also called an inverter, with input A and output Y = A where inversion, denoted by the prime, is defined by the truth table in Fig (b). The value of the output Y = A is the opposite of the input A; that is, A = 1 when A = 0 and A = 0 when A = 1. We emphasize that a NOT gate can have only one input, whereas the OR andand gates may have two or more inputs. Fig

12 CHAP. 15] BOOLEAN ALGEBRA 379 Suppose, for instance, a NOT gate is asked to process the following three sequences: A 1 = , A 2 = , A 3 = The NOT gate changes 0 to 1 and 1 to 0. Thus A 1 = , A 2 = , A 3 = are the three corresponding outputs. Logic Circuits A logic circuit L is a well-formed structure whose elementary components are the above OR, AND, and NOT gates. Figure is an example of a logic circuit with inputs A, B, C and output Y. A dot indicates a place where the input line splits so that its bit signal is sent in more than one direction. (Frequently, for notational convenience, we may omit the word from the interior of the gate symbol.) Working from left to right, we express Y in terms of the inputs A, B, C as follows. The output of the AND gate is A B, which is then negated to yield (A B). The output of the lower OR gate is A + C, which is then negated to yield (A + C). The output of the OR gate on the right, with inputs (A B) and (A + C), gives us our desired representation, that is, Y = (A B) + (A + C) Fig Logic Circuits as a Boolean Algebra Observe that the truth tables for the OR, AND, and NOT gates are respectively identical to the truth tables for the propositions p q (disjunction, p or q ), p q (conjunction, p and q ), and p (negation, not p ), which appear in Section 4.3. The only difference is that 1 and 0 are used instead of T and F. Thus the logic circuits satisfy the same laws as do propositions and hence they form a Boolean algebra. We state this result formally. Theorem 15.12: Logic circuits form a Boolean Algebra. Accordingly, all terms used with Boolean algebras, such as, complements, literals, fundamental products, minterms, sum-of-products, and complete sum-of-products, may also be used with our logic circuits. AND-OR Circuits The logic circuit L which corresponds to a Boolean sum-of-products expression is called an AND-OR circuit. Such a circuit L has several inputs, where: (1) Some of the inputs or their complements are fed into each AND gate. (2) The outputs of all the AND gates are fed into a single OR gate. (3) The output of the OR gate is the output for the circuit L. The following illustrates this type of a logic circuit.

13 380 BOOLEAN ALGEBRA [CHAP. 15 EXAMPLE Figure is a typical AND-OR circuit with three inputs, A, B, C and output Y. We can easily express Y as a Boolean expression in the inputs A, B, C as follows. First we find the output of each AND gate: (a) The inputs of the first AND gate are A, B, C; hence A B C is the output. (b) The inputs of the second AND gate are A, B, C; hence A B C is the output. (c) The inputs of the third AND gate are A and B; hence A B is the output. Then the sum of the outputs of the AND gates is the output of the OR gate, which is the output Y of the circuit. Thus: Y = A B C + A B C + A B Fig NAND and NOR Gates There are two additional gates which are equivalent to combinations of the above basic gates. (a) A NAND gate, pictured in Fig (a), is equivalent to an AND gate followed by a NOT gate. (b) A NOR gate, pictured in Fig (b), is equivalent to an OR gate followed by a NOT gate. The truth tables for these gates (using two inputs A and B) appear in Fig (c). The NAND and NOR gates can actually have two or more inputs just like the corresponding AND and OR gates. Furthermore, the output of a NAND gate is 0 if and only if all the inputs are 1, and the output of a NOR gate is 1 if and only if all the inputs are 0. Fig Observe that the only difference between the AND and NAND gates between the OR and NOR gates is that the NAND and NOR gates are each followed by a circle. Some texts also use such a small circle to indicate a complement before a gate. For example, the Boolean expressions corresponding to two logic circuits in Fig are as follows: (a) Y = (A B), (b) Y = (A + B + C)

14 CHAP. 15] BOOLEAN ALGEBRA 381 Fig TRUTH TABLES, BOOLEAN FUNCTIONS Consider a logic circuit L with n = 3 input devices A, B, C and output Y, say Y = A B C + A B C + A B Each assignment of a set of three bits to the inputs A, B, C yields an output bit for Y. All together there are 2 n = 2 3 = 8 possible ways to assign bits to the inputs as follows: 000, 001, 010, 011, 100, 101, 110, 111 The assumption is that the sequence of first bits is assigned to A, the sequence of second bits to B, and the sequence of third bits to C. Thus the above set of inputs may be rewritten in the form A = , B = , C = We emphasize that these three 2 n = 8-bit sequences contain the eight possible combinations of the input bits. The truth table T = T (L) of the above circuit L consists of the output sequence Y that corresponds to the input sequences A, B, C. This truth table T may be expressed using fractional or relational notation, that is, T may be written in the form T(A,B,C)= Y or T (L) =[A, B, C; Y ] This form for the truth table for L is essentially the same as the truth table for a proposition discussed in Section 4.4. The only difference is that here the values for A, B, C, and Y are written horizontally, whereas in Section 4.4 they are written vertically. Consider a logic circuit L with n input devices.there are many ways to form n input sequences A 1,A 2,...,A n so that they contain the 2 n different possible combinations of the input bits. (Note that each sequence must contain 2 n bits.) One assignment scheme is as follows: A 1 : Assign 2 n 1 bits which are 0 s followed by 2 n 1 bits which are 1 s. A 2 : Repeatedly assign 2 n 2 bits which are 0 s followed by 2 n 2 bits which are 1 s. A 3 : Repeatedly assign 2 n 3 bits which are 0 s followed by 2 n 3 bits which are 1 s. And so on. The sequences obtained in this way will be called special sequences. Replacing 0 by 1 and 1 by 0 in the special sequences yields the complements of the special sequences. Remark: Assuming the input are the special sequences, we frequently do not need to distinguish between the truth table and the output Y itself. T (L) =[A 1,A 2,...,A n ;Y ]

15 382 BOOLEAN ALGEBRA [CHAP. 15 EXAMPLE (a) Suppose a logic circuit L has n = 4 input devices A, B, C, D. The 2 n = 2 4 = 16-bit special sequences for A, B, C, D follow: That is: A = , C = B = , D = (1) A begins with eight 0 s followed by eight 1 s. (Here 2 n 1 = 2 3 = 8.) (2) B begins with four 0 s followed by four 1 s, and so on. (Here 2 n 2 = 2 2 = 4.) (3) C begins with two 0 s followed by two 1 s, and so on. (Here 2 n 3 = 2 1 = 2.) (4) D begins with one 0 followed by one 1, and so on. (Here 2 n 4 = 2 0 = 1.) (b) Suppose a logic circuit L has n = 3 input devices A, B, C. The 2 n = 2 3 = 8-bit special sequences for A, B, C and their complements A, B, C are as follows: A = , B = , C = A = , B = , C = Figure contains a three-step algorithm for finding the truth table for a logic circuit L where the output Y is given by a Boolean sum-of-products expression in the inputs. Fig EXAMPLE Algorithm 15.5 is used to find the truth table T = T (L) of the logic circuit L in Fig or, equivalently, of the above Boolean sum-of-products expression Y = A B C + A B C + A B (1) The special sequences and their complements appear in Example 15.14(b). (2) The products are as follows: A B C = , A B C = , A B = (3) The sum is Y = Accordingly, T( , , ) = or simply T (L) = where we assume the input consists of the special sequences.