InternationalJournalonElectricalEngineeringandInformaticsVolume4,Number1,March2012

Differential Evolution and Genetic Algorithm for Sidelobe Reduction of A Concentric Ring Array Antenna by Radial Variation of Amplitudes With Fixed and Variable First Null Beamwidth
Anirban Chatterjee and G.K. Mahanti Department of Electronics and Communication Engineering, National Institute of Technology, Durgapur, India [email protected], [email protected] Abstract: Reduction of side lobe levels in a concentric ring arrays results in wide first null beamwidth. The authors propose a pattern synthesis method based on the Differential Evolution (DE) algorithm to reduce the side lobe levels of a concentric ring array of isotropic antennas while keeping the first null beamwidth (FNBW) fixed and variable by radial variation of amplitudes of the array elements. Two different cases have been studied, one with fixed initial interelement distance and another with optimum interelement distance for the entire array. The FNBW of the optimized array is kept equal to or less than that of a uniformly excited and 0.5 spaced concentric ring array of the same number of elements and rings. Results are also compared with Genetic Algorithm to establish its superiority. Keywords: Circular ring array, Differential Evolution (DE) algorithm, first null beam width (FNBW), optimization, side lobe level, Genetic algorithm. 1. Introduction A circular ring array, also known as a concentric circular array (CCA) is a planar array that consists of one or more concentric rings, each having equally spaced array elements on its circumference. An important property of a CCA is that, its beam pattern remains invariant for 360 azimuthal coverage if the array consists of several rings with an appreciably large number of elements in each ring. Its main attraction is the cylindrical symmetry of its radiation pattern and compact structure. One of the important configurations regarding CCA is the uniform concentric circular array (UCCA) where the inter-element spacing in each individual ring is kept almost half of the wavelength and all the elements in the array are uniformly excited. However, in its modest form the array suffers from a high side lobe problem. Generally low side lobes in the array factor are obtained through optimum amplitude weights of the signals at each array element. The radiation pattern function of a concentric ring array has been expressed by Stearns and Stewart [1] as a truncated Fourier-Bessel series and the non uniform distribution of the rings has been approximated to a smaller number of equally spaced ones. N. Goto and D. K Cheng showed that for a Taylor weighted ring array the maximum allowable inter-element spacing should be about four-tenths of a wavelength, if high side lobes are to be avoided [2]. L. Biller and G. Friedman used steepest descent iterative process to find out element weights and ring spacing to get lower sidelobe levels and control over beam width [3]. D. Huebner reduced the sidelobe levels for small concentric ring array by adjusting the ring radii using optimization technique [4]. B. P. Kumar and G. R. Branner also proposed optimum ring radii for getting lower sidelobes [5]. M. Dessouky, H. Sharshar and Y. Albagory showed that the existence of central element in concentric circular array of smaller innermost ring reduced the sidelobe levels significantly while minor increase in the beamwidth [6]. Sidelobe levels can be reduced by thinning the array [7-8]. The array is thinned by turning off selected elements from the

Received:May 30th,2011.Accepted:November24th,2011
1

Anirban Chatterjee, et al.

uniform array. Sidelobe level can also be reduced by optimizing both radii of the rings and the number of elements in each ring of a concentric ring array. In this problem, the sidelobe level of a concentric ring array have been reduced by finding out an optimum set of radial amplitude distribution of the array using Differential Evolution (DE) algorithm. Two different cases have been studied. In the first case, the sidelobe level has been reduced by computing the optimum radial amplitude of the array while considering a fixed interelement distance for the entire array. In the second case, the sidelobe level has been reduced by finding out an optimum set of radial amplitude along with an optimum interelement distance for the entire array. In each of the individual cases, the array has been optimized for two different design consideration of fixed first null beamwidth (FNBW) and variable first null beamwidth (FNBW). The main purpose of the optimum interelement spacing for the entire array is to reduce the number of elements in the array. Here Differential Evolution (DE) has been successfully applied as an evolutionary algorithm [9-10] to find out these optimum values. One of the advantages of DE is that it uses a few control parameters. Differential Evolution has been applied as an evolutionary algorithm in various types of antenna array synthesis problems [15-18]. In this problem Differential Evolution (DE) [11-18] has been presented as an efficient alternative to other optimization algorithms like Genetic Algorithm (GA) [19-20], Particle Swarm Optimization (PSO) algorithm [21-22] etc. in handling certain kind of optimization problem. The paper is organized as follows: Section 2 explains the methods of side lobe reduction; Section 3 presents the Differential Evolution algorithm; Section 4 presents the simulation results; and finally the conclusion is presented in Section 5.

Figure1. Concentric ring arrays of isotropic antennas in XY plane 2. Methods of Side lobe Reduction The far field pattern of a concentric circular planar array [6] shown in Figure1 on the x y plane with central element feeding can be defined as:

E( , ) = 1 +

m =1 n =1

I me j[ k rm sin cos( mn )+m ]

M Nm

(1)

Normalized absolute power pattern, P (,) in dB can be expressed as follows:

P( , ) = 10 log 10

E( , ) E( , ) max

(2)

Differential Evolution and Genetic Algorithm for Sidelobe Reduction of A Concentric Ring Array

where, M = Number of concentric rings, Nm = Number of isotropic elements in m-th ring, Im = excitation amplitude of elements on m-th circular ring, dm = inter-element arc spacing of m-th circle, rm = Nmdm/2 is the radius of the mth ring, mn = 2n / N m is the angular position of mn-th element with 1 n Nm , , = polar and azimuth angle, = wave length, k = wave number = 2/, j=complex number, m = excitation phase of elements on m-th ring, All the elements have same excitation phase of zero degree. Side lobe levels of a uniform concentric ring array can be reduced by finding out a suitable set of radial amplitude distribution of the array elements, which is based on the assumption that all the array elements on the same circle have same amplitude distribution, but they vary from ring to ring. The number of elements in m-th ring of a concentric ring array can be expressed as:

Nm =

2 rm dm

(3)

In this problem, two different cases have been considered. For the first case, the interelement distance dm for the entire array is kept at 0.5. Then an optimum set of radial amplitude distribution for the entire array is determined using DE such that the optimized array gives lower side lobe levels while retaining other desire array characteristics. In the second case, optimum values of radial amplitudes and optimum interelement distance dm for the entire array are determined using DE algorithm to get lower sidelobe levels with desire array characteristics. In the second case d m is varied in such a way that it lies between

0.5 d m .

The number of elements in each ring is determined using equation (3). Since the number of elements in a particular ring must be an integer quantity, so only the computed integer values of equation (3) are taken The fitness functions for this problem are given by:

Fitness1 = k1 max SLL + k2 ( FNBWo FNBWd ) H( T ) Fitness 2 = max SLL

(4) (5)

Where max SLL is the value of maximum sidelobe level, FNBWo , FNBWd are the obtained and desired values of first null beam width respectively, k1, k2 are weighting coefficients to control the relative importance given to each term of equation (4) and the values are chosen as, k1=1 and k2=100 respectively. H(T) is Heaviside step functions defined as:
T = ( FNBWo FNBWd )

(6)

H(T ) =

0, 1,

if T < 0 , if T 0 (7)

Equation (5) is for not keeping FNBW fixed. Equation (4) and equation (5) are minimized using DE for optimal synthesis of array.
3. Differential Evolution Algorithm Differential Evolution is a simple evolutionary algorithm introduced by Storn and Price [11]. Similar to GA [17, 19-20], DE is also an algorithm based on population. DE algorithm is a stochastic optimization method for minimizing an objective function that can model the problems objectives while incorporating constraints. The algorithm mainly has three

Anirban Chatterjee, et al.

advantages: ability to find the true global minima regardless of the initial parameter value, converges fast and uses a few control parameter [11-14]. DE first samples the objective function at multiple, randomly chosen initial points. Then NP (Number of populations) vectors in the initial population are chosen from the predefined parameter bounds. To explore the objective functions landscape, DE employs a difference between the parameter vectors. New points (trial solution) are generated through perturbations of existing points. DE perturbs vectors with the scale difference of two randomly selected population vectors. In the next stage (selection) the trial vector competes against the population vector of the same index. Once the last trial vector has been tested, the survivors of the NP pair wise competitions become parents for the next generation in the evolutionary cycle. The algorithm can be summarized as below [11-14]: Step 1:Initialization: The generation number is set to t=0 and a population of NP individuals are randomly initialized in the D-dimensional search space as:

r r r X 1 ( t ) , , X NP ( t ) , where X ( t ) = [ xi,1( t ), xi,2 ( t ),..., xi,D ( t )] and i r r , X max ] . each individuals are uniformly distributed in domain [ X min Pt =

Step 2: Evaluate the fitness: Evaluate the fitness of each individual at current generation. Step 3: Mutation: r Create donor vector Vi ( t ) corresponding to the

r i-th target vector X i ( t ) for all the

r r r r for i=1,2,...,NP Vi ( t ) = X best ( t ) + F .( X i ( t ) X i ( t )) r1 r2 r r r is the best vector of the current population, X i and X i are randomly picked where, X best
up vectors from the current generations, F is the scale factor, F ( 0 ,1+ ) , a positive real number that controls the rate at which the population evolves.
Step 4: Crossover: r Use any one of the crossover scheme in DE [14-16] to form the trial vector U i ( t ) , by r r exchanging the components of the donor vector Vi ( t ) and the target vector X i ( t ) with a

individuals at current generation using any one of the DE mutation scheme [10-16]. In this problem the mutation strategy known as DE/best/1 has been used and is expressed as:

r1

r2

crossover probability of Cr ( Cr [ 0 ,1]) , for all the individuals at current generation.

Compute X Gbest (t ) at current generation as follows: Find out the corresponding vector among NP individuals for which

f ( X i (t + 1)), for , i = 1, 2,..., NP , becomes minimum (for minimization problem) and r r assign that vector to X Gbest (t ) where, f ( X ) is the function to be minimized. Since the

Differential Evolution and Genetic Algorithm for Sidelobe Reduction of A Concentric Ring Array

selection process employs a binary decision the population size remains fixed throughout generations.
Step 6: Increase the iteration count t = t+1 and repeat step 2-5 until the termination condition is satisfied. Return X Gbest as the result.

The termination condition can be defined: (i). When a fixed number of iteration for tmax , with a suitably large value of tmax , depending upon the complexity of the objective function, is reached. (ii). When best fitness of the population does not change appreciably over successive iterations. Mutation demarcates one DE scheme from another. Each mutation strategy combines with either exponential or binomial type crossover and produce new working strategy. There are in total ten different working strategies of DE as suggested by Storn and Price [11-14]. In this problem the DE/best/1/exp strategy has been used along with number of population (NP) =40 and crossover rate (CR) = 0.7 and the termination condition has been defined as tmax=800. Number of variables for the first case is = 9 and for the second case is=10. For the first case, the lower limits for nine variables have been taken as: X min = [0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1] and the upper limits are taken as:

X max = [1, 1, 1, 1, 1, 1, 1, 1, 1, 1]. For the second case, the lower limits for ten variables are taken as: X min = [0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.51] and the upper limits are taken as: X max = [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]. In the second case the lower limit of the tenth variable
is taken 0.51, because the optimum value of the interelement spacing for the entire array should be more than 0.5 in order to avoid the effect of mutual coupling.
4. Simulation Results For a nine ring concentric ring array of isotropic antennas [7], the initial radius of the rings

are rm = m (m-th ring) and the interelement spacing in each ring is taken as . For this 2 2 arrangement the total number of isotropic elements is 279. Uniform excitation and constant phase angle between the elements gives side lobe level -17.4 dB [7] and FNBW 14.8 degree. The reason behind choosing concentric ring array is its -symmetric beam pattern and compact structure. In this problem, the sidelobe level of the array has been reduced based on finding out an optimum set of amplitude distribution of the array under two different cases. The obtained results using DE for both the cases are also compared with Genetic Algorithm (GA). The fitness functions for the GA are taken same as DE. The numbers of population in case of GA are taken same as that of DE. Two-point crossover along with uniform mutation of rate 0.01 and ranking selection are used. Crossover fraction is taken to be 0.07. The initial values and the termination conditions are also kept same as DE. Case I: In this case, interelement distance is kept fixed ( d m

= 0.5 ) for the entire array. Total

number of isotropic elements in the array is 279. To get lower side lobe level with FNBW below or equal to that of a nine ring uniform concentric ring array, an optimum set of radial amplitude distribution has been found out using Differential Evolution (DE) algorithm. In this way the side lobe levels has been reduced below -22 dB with fixed FNBW and below -43 dB without fixing FNBW. The obtained values of side lobe level using GA are -21.22 dB for fixed FNBW and -39.82 dB for variable FNBW.

Anirban Chatterjee, et al.

d m is not prefixed but also optimized. d m is varied in such a way that it lies between 0.5 d m .
To obtain lower side lobe level with FNBW below or equal to that of a nine ring uniform concentric ring array, optimum set of radial amplitude distribution and optimum value of interelement spacing for entire array has been found out using Differential Evolution (DE) algorithm. In this way, the side lobe levels have been reduced below -22 dB with fixed FNBW and below -44 dB without fixing FNBW. The obtained values of side lobe level using GA are -21.37 dB for fixed FNBW and -40.11 dB for variable FNBW. In the second case as the computed value of d m is greater than 0.5 , so the total number of elements in the optimized arrays are also reduced. Table 1. Maximum side lobe levels, FNBW and computation time for the optimized arrays with and without fixed FNBW computed individually using DE and GA.
DE Types of array Maximum sidelobe level (dB) FNBW (degree) Time (hr : min) Maximum side lobe level (dB) GA FNBW (degree) Time (hr : min)

Case II: In this case interelement distance

Optimized radial amplitude with fixed dm (Fixed FNBW) Optimized radial amplitude with optimized

-22.077

14.800

2:08

-21.22

14.800

2:17

-22.079

14.800

1:56

-21.37

14.800

2:09

dm
(Fixed FNBW) Optimized radial amplitude with fixed

dm
(Variable FNBW) Optimized radial amplitude with optimized

-43.946

24.800

2:05

-39.82

23.400

2:11

dm
(Variable FNBW)

-44.070

25.000

1:54

-40.11

23.200

2:13

All these simulations are performed using a PC having Intel core2 duo processor with 3 GHz clock frequency, 2 GB of RAM and Microsoft windows XP 32 bit operating system. Table 1 shows the maximum side lobe level, FNBW and computation time for the optimized arrays with and without fixed FNBW computed individually using DE and GA. Table 1 also shows the relative improvements in the side lobe levels when optimum interelement distance is computed along with radial amplitude distributions keeping FNBW fixed or variable. From Table 1 it can be noticed that the array optimized using DE gives better side lobe levels than the optimized array using GA for all the cases.

Differential Evolution and Genetic Algorithm for Sidelobe Reduction of A Concentric Ring Array

Table 2. Radial amplitude, inter element distance and number of elements in each ring in uniform array [7]
Parameter Im Fixed dm() Nm (2rm/dm) Ring 1 1.000 0.500 6 Ring 2 1.000 0.500 12 Ring 3 1.000 0.500 18 Ring 4 1.000 0.500 25 Ring 5 1.000 0.500 31 Ring 6 1.000 0.500 37 Ring 7 1.000 0.500 43 Ring 8 1.000 0.500 50 Ring 9 1.000 0.500 56

Table 3. Radial amplitude, prefixed inter element distance (0.5) and number of elements in each ring in a optimized array with fixed FNBW computed using DE
Parameter Im Nm (2rm/dm) Ring 1 0.4450 6 Ring 2 0.748 12 Ring 3 0.518 18 Ring 4 0.619 25 Ring 5 0.389 31 Ring 6 0.525 37 Ring 7 0.237 43 Ring 8 0.106 50 Ring 9 1.000 56

Table 4. Radial amplitude, prefixed inter element distance (0.5) and number of elements in each ring in an optimized array for variable FNBW computed using DE
Parameter Im Nm (2rm/dm) Ring 1 0.933 6 Ring 2 1.000 12 Ring 3 0.836 18 Ring 4 0.695 25 Ring 5 0.520 31 Ring 6 0.396 37 Ring 7 0.253 43 Ring 8 0.158 50 Ring 9 0.100 56

Table 5. Radial amplitude, prefixed inter element distance (0.5) and number of elements in each ring in a optimized array with fixed FNBW computed using GA
Parameter Im Nm (2rm/dm) Ring 1 0.715 6 Ring 2 0.748 12 Ring 3 0.861 18 Ring 4 0.751 25 Ring 5 0.325 31 Ring 6 0.7665 37 Ring 7 0.1681 43 Ring 8 0.531 50 Ring 9 0.999 56

Table 6. Radial amplitude, prefixed inter element distance (0.5) and number of elements in each ring in an optimized array for variable FNBW computed using GA
Parameter Im Nm (2rm/dm) Ring 1 0.999 6 Ring 2 0.999 12 Ring 3 0.872 18 Ring 4 0.768 25 Ring 5 0.521 31 Ring 6 0.460 37 Ring 7 0.267 43 Ring 8 0.217 50 Ring 9 0.110 56

Anirban Chatterjee, et al.

Table 7. Radial amplitude, optimized inter element distance and number of elements in each ring in an optimized array for fixed FNBW computed using DE
Parameter Im Optimized dm() Nm (2rm/dm) Ring 1 0.3380 0.602 5 Ring 2 0.8100 0.602 10 Ring 3 0.4460 0.602 15 Ring 4 0.6780 0.602 20 Ring 5 0.3530 0.602 26 Ring 6 0.5210 0.602 31 Ring 7 0.2340 0.602 36 Ring 8 0.1000 0.602 41 Ring 9 0.9980 0.602 46

Table 8. Radial amplitude, optimized inter element distance and number of elements in each ring in an optimized array for variable FNBW computed using DE
Parameter Im Optimized dm() Nm (2rm/dm) Ring 1 0.936 0.511 6 Ring 2 1.000 0.511 12 Ring 3 0.838 0.511 18 Ring 4 0.723 0.511 24 Ring 5 0.537 0.511 30 Ring 6 0.405 0.511 36 Ring 7 0.252 0.511 43 Ring 8 0.160 0.511 49 Ring 9 0.100 0.511 55

Table 9. Radial amplitude, optimized inter element distance and number of elements in each ring in an optimized array for fixed FNBW computed using GA
Parameter Im Optimized dm() Nm (2rm/dm) Ring 1 0.656 0.674 4 Ring 2 0.937 0.674 9 Ring 3 0.246 0.674 13 Ring 4 0.993 0.674 18 Ring 5 0.411 0.674 23 Ring 6 0.438 0.674 27 Ring 7 0.319 0.674 32 Ring 8 0.370 0.674 37 Ring 9 0.956 0.674 41

Table 10. Radial amplitude, optimized inter element distance and number of elements in each ring in an optimized array for variable FNBW computed using GA
Parameter Im Optimized dm() Nm (2rm/dm) Ring 1 0.8637 0.628 5 Ring 2 0.999 0.628 10 Ring 3 0.858 0.628 15 Ring 4 0.723 0.628 20 Ring 5 0.582 0.628 25 Ring 6 0.433 0.628 30 Ring 7 0.282 0.628 35 Ring 8 0.205 0.628 40 Ring 9 0.140 0.628 45

Differential Evolution and Genetic Algorithm for Sidelobe Reduction of A Concentric Ring Array

Figure 2. A nine-ring concentric ring array in the x-y plane of 279 isotropic elements and interelement spacing of 0.5 Table 2 shows radial amplitude distribution, inter-element spacing and number of elements in each ring for a uniformly excited circular ring array. The amplitude of the central element of the array is always kept 1 for all the cases. Table 3 and Table 4 shows optimized radial amplitude distributions computed using DE when interelement distance is kept fixed at 0.5 for fixed and variable FNBW. Table 5 and Table 6 shows optimum results for the same conditions computed using GA.

Figure 3. Normalized absolute power pattern in dB for the uniformly excited concentric ring array of nine concentric rings and 279 isotropic elements.

Anirban Chatterjee, et al.

Figure 4. Normalized absolute power patterns in dB for optimized arrays of fixed interelement distance keeping fixed FNBW computed individually using DE and GA Table 7 shows optimum radial amplitude distributions and optimum interelement distance computed using DE for fixed FNBW whereas Table 8 shows the same parameters for variable FNBW. Table 9 and Table 10 shows the optimum results for the same conditions computed using GA. From Table 7, Table 8, Table 9 and Table 10, it can also be noticed that the optimum values of the interelement spacing for keeping FNBW fixed or variable is greater than 0.5 . So the total number of elements in the optimized arrays of second case is lesser than that of optimized arrays of first case having fixed interelement distance d m = 0.5 . Figure 2 shows the uniformly excited nine-ring concentric ring array in the x-y plane. The normalized power pattern of the array is shown in Figure 3. Figure 4 and Figure 5 shows the normalized power patterns in dB for the optimized arrays of fixed interelement spacing with fixed and variable FNBW computed using DE and GA.

Figure 5. Normalized absolute power patterns in dB for optimized arrays of fixed interelement distance keeping variable FNBW computed individually using DE and GA

10

Differential Evolution and Genetic Algorithm for Sidelobe Reduction of A Concentric Ring Array

Figure 6. Normalized absolute power patterns in dB for optimized arrays with optimized interelement distance keeping fixed FNBW computed individually using DE and GA

Figure 7. Normalized absolute power patterns in dB for optimized arrays with optimized interelement distance for variable FNBW computed individually using DE and GA Figure 6 and Figure 7 shows the normalized power patterns in dB for the optimized arrays having optimized interelement spacing with fixed FNBW and variable FNBW computed individually using DE and GA. The convergence characteristics of DE and GA for both the cases are shown in Figure8 and Figure9 respectively. From Figure8 and Figure9 it can be observed that the performance of DE is better than GA while reducing the fitness functions under case I and case II.

11

Anirban Chatterjee, et al.

Figure 8. Convergence characteristics of DE and GA for case I: (a) with fixed FNBW consideration (b) with variable FNBW consideration.

Figure 9. Convergence characteristics of DE and GA for case II: (a) with fixed FNBW consideration (b) with variable FNBW consideration.
Conclusions The paper presents a method based on Differential Evolution algorithm to reduce the side lobe level of a concentric ring array antenna by finding out an optimum set of radial amplitude distribution of the array. Two different cases comprising of fixed and optimized inter-element spacing for the array has been studied. In each of the individual cases, the array is optimized separately for two different design considerations of fixed and variable FNBW. Results clearly show a very good agreement between desired and obtained value. Here Differential Evolution (DE) has been effectively used as a global optimization algorithm to find out the optimum set of radial amplitudes for the first case and find out optimum set of radial amplitudes with optimum inter-element distance for the second case. Results are also compared with Genetic Algorithm to establish its superiority. It can also be used for synthesizing other array configurations.

12

Differential Evolution and Genetic Algorithm for Sidelobe Reduction of A Concentric Ring Array

References [1] Stearns, C., and A. Stewart, An investigation of concentric ring antennas with low sidelobes, IEEE Trans. Antennas Propag., Vol. 13, No. 6, 856863, 1965. [2] Goto, N., and D. K. Cheng, On the synthesis of concentric-ring arrays, IEEE Proc., Vol. 58, No. 5, 839840, 1970, [3] Biller, L., and Friedman, G. (1973), Optimization of radiation patterns for an array of concentric ring sources, IEEE Trans. Audio Electroacoust., Vol. 21, No. 1, 5761, 1973. [4] Huebner, D.,Design and optimization of small concentric ring array, Proc. IEEE Antennas Propagation Int. Symp., Vol. 16, 455458, 1978. [5] Kumar, B. P., and G. R. Branner, Design of low sidelobe circular ring array by element radius optimazation, Proc. IEEE Antennas Propagation Int. Symp., 20322035, 1999. [6] Dessouky, M., H. Sharshar, and Y. Albagory, Efficient sidelobe reduction technique for small-sized concentric circular array, Progress In Electromagnetics Research, PIER 65, 187200, 2006. [7] Haupt, R. L., Optimized Element Spacing for Low Sidelobe Concentric Ring Array, IEEE Transactions On Antennas and Propagation, Vol. 56, No. 1, 266-268, 2008. [8] Haupt, R. L. Thinned concentric ring array, IEE proc., Antennas and Propagation Int. Symp., 1-4, 2008. [9] David B. Fogel., Evolutionary computation: toward a new philosophy of machine intelligence, 3rd Edition, John Wiley and Sons, 2006. [10] Thomas Baeck, D. B. Fogel, and Z Michalewicz., Handbook of Evolutionary Computation, Taylor & Francis, New York, 1997. [11] Storn, R. and K. Price, Differential evolutionA simple and efficient heuristic for global optimization over continuous spaces, Journal of Global Optimization, Vol. 11, No. 4, 341-359, 1997. [12] Kenneth V. Price, Rainer M. Storn, Jouni A. Lampinen, Differential evolution: a practical approach to global optimization, Springer, 2005. [13] Vitaliy Feoktistov , Differential Evolution: In Search of Solutions, Springer, 2006. [14] Swagatam Das, Ajith Abraham and Amit Konar, Metaheuristic Clustering, Springer, 2009. [15] Yang, S. A. Qing, Y. B. Gan, Synthesis of low sidelobe antenna arrays using the differential evolution algorithm, IEEE Proc, Antennas and Propagation Society International Symposium, Vol.1, 780 - 783 , 2003 [16] Rocha-Alicano, C., D. Covarrubias-Rosales, C. Brizuela-Rodriguez, and M. PanduroMendoza, Differential evolution algorithm applied to sidelobe level reduction on a planar array, AEU International Journal of Electronic and Communications, Vol. 61, No. 5, 286-290, 2007. [17] Panduro, M. A., C. A. Brizuela, L. I. Balderas, and D. A. Acosta, A comparison of genetic algorithms, particle swarm optimization and the differential evolution method for the design of scannable circular antenna arrays, Progress In Electromagnetics Research B, Vol. 13, 171-186, 2009. [18] Pal, S., B. Qu, S. Das, and P. N. Suganthan, Linear antenna array synthesis with constrained multi-objective differential evolution, Progress In Electromagnetics Research B, Vol. 21, 87-111, 2010 [19] Man, K.F., K.S. Tang and S Kwong., Genetic Algorithms: Concepts and Applications, IEEE Transactions on Industrial Electronics, Vol.43, Issue: 5, 519 534, 1996. [20] Mahanti, G.K., A .Chakraborty, and S. Das. Design of Phase-Differentiated Reconfigurable Array Antennas with Minimum Dynamic Range Ratio, IEEE Antennas and Wireless Propagation Letters, Vol.5, No. 1, 262-264, 2006. [21] Kennedy, J. and R. Eberhart, Particle swarm optimization, IEEE Proc., International Conference Neural Networks, Vol. 4, 1942-1948, 1995.

13

Anirban Chatterjee, et al.

[22] Jin, N., and Y. Rahmat-Samii, Advances in particle swar optimizatio for antenna R n rm on a gns: Real-num mber, binary, single- objecti ive and multi iobjective imp plementations, desig IEEE Trans. Anten E nnas Propag., V 55, No. 3, 556-567, 2007. Vol. , Anirban Chatterje received his Bachelor of T ee s Technology in Electronics & munication En ngineering in the year of 2006 from D B.C. Roy Dr. y Comm Engin neering Colle ege, Durgapur India and Master of T r, Technology in n Telec communication Engineering in 2009 f n g from National Institute of l f Techn nology, Durga apur, India. His research inte erest includes A Array Antenna a synth hesis, Soft Com mputing & Elec ctromagnetics. He has publis shed 6 research h paper in internatio rs onal journals a 5 papers i internationa conferences. and in al . Curre ently he is wor rking towards his PhD degre from Nation Institute of ee nal f Technology, Durgapur, India.

G. K. Mahanti was born in the ye 1967 in We Bengal, Indi He obtained . s ear est ia. d his B.E. in Ele ectronics & Communicatio Engg. in 1988 from on n m NIT,D Durgapur,India M.E. in Elect a, tronics System and Communi m ication in 1994 4 from NIT,Rourkela, ,India and Ph.D D(Engg.) from IIT,Kharagpur r,India. He has s rs Presently he is s more than 20 year of teaching and research experience. P ing as Prof fessor and H Head, departm ment of Ele ectronics and d worki Comm munication En ngineering, Na ational Institute of Technolo e ogy, Durgapur, , India. He has publis shed more than 40 papers in journals and i national and n n in d es. e many internatio onal journals li Electronics ike s international conference He was the reviewer of m EEE a tter, Progress in Electromagne n etics Research, , Letter, IE Antennas and Wireless Propagation Let Internatio onal Journal of Adaptive con f ntrol & Signal processing and many confer d rences. He was s also the program comm mittee membe of many na er ational and in nternational con nferences. His s biography is listed in Marquis Who is Who in the world. His research area is Array antenna y M a synthesis, Evolutionary Algorithms &E , Electromagneti ics.

14