International Journal of VLSI design & Communication Systems (VLSICS) Vol.4, No.

4, August 2013

EXTENDED K-MAP FOR MINIMIZING MULTIPLE OUTPUT LOGIC CIRCUITS

Palash Das1, Bikromadittya Mondal2
Department of Computer Science and Technology, Bengal Engineering and Science University, Shibpur, Howrah, India. 2 Department of Computer Science and Engineering, B P Poddar Institute of Management and Technology, Kolkata, India.
1

ABSTRACT
Minimization of multiple output functions of a digital logic circuit is a classic research problem. Minimal circuit is obtained by using multiple Karnaugh Maps (K-map), one for each function. In this paper we propose a novel technique that uses a single Karnaugh Map for minimizing multiple outputs of a single circuit. The algorithm basically accumulates multiple K-Maps into a single K-Map. Finding minimal numbers of minterms are easier using our proposed clustering technique. Experimental results show that minimization of digital circuits where more than one output functions are involved, our extended K-Map approach is more efficient as compare to multiple K-Map approach.

KEYWORDS
Boolean Algebra, Karnaugh Map, Digital Logic Circuit, Clustering.

1. INTRODUCTION
Simplification of logic function actually reduces the number of digital logic gates required to implement digital circuits. This results the reduction of the size of the circuit. The cost of the circuit will also be reduced. There are a number of techniques proposed to minimize logic functions. The Boolean algebra was proposed by Boole [1]. Then C.E. Shannon [2] showed the design of digital circuits using Boolean algebra. Karnaugh [3] proposed a new technique for simplifying Boolean expressions using a map. Quine and McCluskey [4][5] proposed an algorithmic based technique for simplifying Boolean logic functions. S. K. Petrick [6] also did significant work on Boolean function minimization. Heuristic based techniques [9][10] were proposed for fast minimization of Boolean functions. Generally Boolean functions are expressed in terms of two standard forms: the sumofproducts and the productofsums. Each combination of variables in a sum-of-products function is called a minterm; in the product-of-sums form, they are called maxterms. This paper presents the use of minterms to create Extended Karnaugh Map (K-map), although the same technique can also be used for creating Extended K-map for Boolean functions by maxterm expressions. This Extended K-map can accommodate more than one output functions at a time and helps to design the entire minimized circuit at a time. It is obvious that the number of variables of all the functions will be same as this new technique of Extended K-map has been developed to design specific circuits of a particular instance of time. If there are m number of n variable output functions then this Extended K-map will have 2n cells which will accommodate all m functions and will produce
DOI : 10.5121/vlsic.2013.4401 1

International Journal of VLSI design & Communication Systems (VLSICS) Vol.4, No.4, August 2013

minimized Boolean expressions for designing the entire circuit. Figure 1 shows the block diagram of the system. It is efficient as it reduces the design complexity significantly in case of multioutput circuits.

Figure 1: Block diagram of an Extended K-map

2. PROPOSED WORK
This work is actually the extended version of K-map which is based on basic K-map principle but with some additional features. The Extended K-map algorithm is presented in section 2.1. The cluster generation and selection algorithm is given in section 2.2. The illustration of our proposed algorithm is given in section 2.3.

2.1 The Extended K-map Algorithm

Input: k number of n variable Boolean functions. Output: k number of minimized Boolean expressions for designing the circuit. Step 1: Draw a map (matrix) of 2n cells; where n is the number of variables of all functions Step 2: Step 2.1: Initialize all variables Step 2.2: Set array_functions = all output functions to be minimized, NO_OF_FUNCTIONS = length of the array_functions Step 3: REPEAT the following steps till NO_OF_FUNCTIONS! = 0 Step 3.1: generate_cluster will return the clusters one by one of the function passed as parameter. Step 3.2: All the returned clusters will in taken into a stack. And later they will be taken out in last in first out (LIFO) basis for getting the minimized outputs Step 3.3: NO_OF_FUNCTIONS; (decrementing the number of remaining functions to be minimized) Step 4: END OF LOOP Step 5: Step 5.1: REPEAT until the stack is empty Step 5.2: Now pop the first cluster Step 5.3: Find the elements (groups) of the cluster and form the minimize Boolean expressions(in SOP or POS form) using basic K-Map principle to design the minimized circuit

International Journal of VLSI design & Communication Systems (VLSICS) Vol.4, No.4, August 2013

2.2 The Clustering Algorithm

This algorithm actually describes the process of making cluster of the corresponding output functions that has been passed as parameter. It returns all the clusters one by one and stores in to the stack. The array_functions is basically a pointer to an array which holds all the minterm or maxterm values of the corresponding functions passed as argument to the process. We are considering SOP forms for this process assuming the POS form of the algorithm can be easily changed as like our basic K-map principle. Step 1: In the first iteration initialize the empty cell by putting 1s on top in the cell for the corresponding minterms of the first function. Step 1.1: Make initial groups of those 1s using basic K-map principle Step 1.2: Assign a name to each group (e.g. X1, X2 etc.) Step 1.3: Take all those groups of the first function into a cluster (e.g. C1={X1, X2}) Step 2: In the other iteration put 1s directly on the cells if the cells are still empty after putting the value of previous functions Step 3: If the cell is not empty; check the previous symbol Step 3.1: If the previous symbol of the cell is 1, make it dont care by putting X just below the 1 Step 3.2: If the previous symbol of the cell is X, make it dont care by putting 1 just below the 1 Step 3.3: Make groups using basic K-map principle of the uncovered 1s and the value of the cell just changed from 1 to X or X to 1 Step 3.4: Assign a name to each of those groups (e.g. X3, X4 etc.) Step 3.5: Take all those groups of each function into a cluster (e.g. C2={X1, X3, X4}) Step 4: Return the clusters one by one to the top of the stack

2.3 Example
Let us consider the following four 4-variable functions for minimization. F1 = ABCD + ABCD + ABCD + ABCD + ABCD + ABCD + ABCD F2 = ABCD + ABCD + ABCD + ABCD F3 = ABCD + ABCD + ABCD + ABCD F4 = ABCD + ABCD + ABCD + ABCD + ABCD + ABCD + ABCD + ABCD

International Journal of VLSI design & Communication Systems (VLSICS) Vol.4, No.4, August 2013

Figure 2 shows the Extended K-map that holds all the four functions. The grouping is done based on our Extended k-map algorithm.

Figure 3: Stack of clusters 4

International Journal of VLSI design & Communication Systems (VLSICS) Vol.4, No.4, August 2013

From the stack we will get clusters one by one which are nothing but the minimized forms of the corresponding functions. C1= {X1, X2} C2= {X5, X6, X7} C3= {X3} C4= {X4} NO_OF_FUNCTIONS -- OUT1 = AB + CD NO_OF_FUNCTIONS --OUT2 = B NO_OF_FUNCTIONS --OUT3 = BCD + ABD +ABCD NO_OF_FUNCTIONS --OUT4 = BD

3. EXPERIMENTAL RESULTS AND SIMULATIONS

Table 1 shows the experimental results of our proposed technique for different circuits.
Table 1: Experimental results

Number of variables
2

Functions under experiment

1. 2. F1(A,B) = E (1, 2) F2(A,B) = E (1, 3)

Minimized output using Extended K-Map

1. 2. K1 = B K2 = A

Reference Extended KMap

Figure 4

1. 2. 3.

F1 = E1 (0, 1, 2, 3, 4, 6) F2 = E2 (3, 5, 6, 7) F3 = E3 (0, 2, 4, 6)

1. 2. 3.

K1 = A + C K2 = AB+BC+CA K3 = C

Figure 5

1. 4 2. 3. 4.

F1 = E1 (3, 7, 11, 12, 13, 14, 15) F2 = E2 (0, 2, 8, 15) F3 = E3 (0, 2, 8, 10) F4 = E4 (0, 1, 2, 3, 8, 9, 10, 11)

1. 2. 3. 4.

K1 = AB + CD K2 = BCD+ ABD +ABCD K3 = BD K4 = B

Figure 6

1. 5 2. 3. 4. 5.

F1 = E1 (1, 2, 3, 5, 7,11, 13, 17, 19, 23, 29, 31) F2 = E2 (8,9,13,12) F3 = E3 (0, 1, 5, 4) F4 = E4 (8,12) F5 = E5 (24, 25, 27, 26, 30, 31, 29, 28)

1.

2. 3. 4. 5.

K1 = ABE + BCE + ABC D +ABDE +ACDE+ ABCE+ACDE K2 = AB'D K3 = ABD K4 = ABDE K5 = AB

Figure 7

International Journal of VLSI design & Communication Systems (VLSICS) Vol.4, No.4, August 2013

Figure 4, Figure 5, Figure 6 and Figure 7 show the minimization process of different multiple output circuits using the Extended K-map.

4. COMPLEXITY ANALYSIS
Generally in case of basic K-Map [3] for n-variable function it needs 2n spaces for a single function to be solved. So for k numbers of functions total number of spaces are k*2n. So space complexity function will be Fk = k * 2n which is O (2n). But in case of our Extended K-Map algorithm for n-variable function it needs 2n spaces for all k function with an additional stack which has again k spaces for k number of functions. So space complexity function will be Fk = k + 2n which is O (2n). So our algorithm is more space efficient as compared with previous one. Table 2 and Figure 8 show the complexity of our Extended K-map method compared with the complexity of basic Kmap method.

International Journal of VLSI design & Communication Systems (VLSICS) Vol.4, No.4, August 2013 Table 2: Complexity comparison

Figure 8: Graphical representation of complexity comparison 7

International Journal of VLSI design & Communication Systems (VLSICS) Vol.4, No.4, August 2013

5. CONCLUSIONS
In this paper, we have proposed a new K-map which is more space efficient with increasing number of functions. Any number of functions can be minimized using this new technique. Future work may be done to extend this technique for minimizing large circuits.

REFERENCES
Boole G. (1954): An Investigation of the Laws of Thought. New York: Dover Publications. Shannon C.E. (1938): A symbolic analysis of relay and switching circuits. Trans. AIEE, Vol. 57, No. 6, pp. 713723. [3] Karnaugh M. (1953): The map method for synthesis of combinatorial logic circuits. Trans. AIEE Comm. Electron.,Vol. 72, No. 4, pp. 593598. [4] McCluskey E. J. (1956), Minimization of Boolean functions, Bell System Tech. J., Vol. 35, No. 5, pp. 14171444. [5] Quine W. V. (1952), The problem of simplifying truth tables, Amer. Math. Month., Vol. 59, No. 8, pp. 521531. [6] Petrick S. K. (1959), On the minimization of Boolean functions, Proc. Int. Conf. Information Processing, Paris: Unesco, pp. 422423. [7] McCluskey E. J. (1965), Introduction to the Theory of Switching Circuits, New York, McGrawHill. [8] Biswas N. N. (1971), Minimization of Boolean Functions, IEEE Trans. on Computers, Vol. C-20, pp. 925-929. [9] Hong S. J., Cain R. G., Ostapko D. L. (1974), MINI: A Heuristic Approach for Logic Minimization, IBM Journal of Research and Development, Vol. 18, pp. 443-458. [10] Rhyne V. T., Noe P. S., McKinney M. H., and Pooch U.W. (1977) A New Technique for the Fast Minimization of Switching Functions, IEEE Trans. on Computers, Vol. C-26, pp. 757-764. [1] [2]