OPTIMIZATION OF QOS PARAMETERS IN COGNITIVE RADIO USING ADAPTIVE GENETIC ALGORITHM

International Journal of Next-Generation Networks (IJNGN) Vol.4, No.2, June 2012
DOI : 10.5121/ijngn.2012.4201 1
OPTIMIZATION OF QOS PARAMETERS IN
COGNITIVE RADIO USING ADAPTIVE GENETIC
ALGORITHM
Maninder Jeet Kaur1
, Moin Uddin2
and Harsh K Verma3
1,3
Department of Computer Engineering, Dr B R Ambedkar National Institute of
Technology, Jalandhar, India
mani356@gmail.com , vermah@nitj.ac.in
2
Delhi Technological University, New Delhi, India
prof_moin@yahoo.com
ABSTRACT
Genetic algorithm based optimization rely on explicit relationships between parameters, observations and
criteria. GA based optimization when done in cognitive radio can provide a criteria to accommodate the
secondary users in best possible space in the spectrum by interacting with the dynamic radio environment
at real time. In this paper we have proposed adaptive genetic algorithm with adapting crossover and
mutation parameters for the reasoning engine in cognitive radio to obtain the optimum radio
configurations. This method ensure better controlling of the algorithm parameters and hence the increasing
the performance. The main advantage of genetic algorithm over other soft computing techniques is its multi
– objective handling capability. We focus on spectrum management with a hypothesis that inputs are
provided by either sensing information from the radio environment or the secondary user. Also the QoS
requirements condition is also specified in the hypothesis. The cognitive radio will sense the radio
frequency parameter from the environment and the reasoning engine in the cognitive radio will take the
required decisions in order to provide new spectrum allocation as demanded by the user. The transmission
parameters which can be taken into consideration are modulation method, bandwidth, data rate, symbol
rate, power consumption etc. We simulated cognitive radio engine which is driven by genetic algorithm to
determine the optimal set of radio transmission parameters. We have fitness objectives to guide one system
to an optimal state. These objectives are combined to one multi – objective fitness function using weighted
sum approach so that each objective can be represented by a rank which represents the importance of each
objective. We have transmission parameters as decision variables and environmental parameters are used
as inputs to the objective function. We have compared the proposed adaptive genetic algorithm (AGA) with
conventional genetic algorithm (CGA) with same set of conditions. MATLAB simulations were used to
analyze the scenarios.
KEYWORDS
Cognitive Radio, Genetic Algorithm, Optimization
1. INTRODUCTION
Because of the growth of the number of services, the spectrum which is a limited natural resource
is becoming scarce. Various studies and measurements show that the spectrum is not being

2
utilized efficiently at any given time and space [1]. It has been noticed that the peak traffic of
each system occurs during specific times and within specific locations. There have been many
studies on the challenges associated with the realization of cognitive radio, some of these are (1)
efficient use of the available spectrum (spectrum holes) (2) Spectrum Mobility (3) Energy
efficiency (4) Spectrum Decision (5) Accurate Sensing and (6) QoS provisioning. The support of
quality of service in optimization of the cognitive radio scenario is also a challenging problem.
The key question is “How the secondary user would find (in the fastest way) about the free slot of
the available spectrum and make necessary decisions to occupy it and at the same time guarantee
the QoS of the primary as well as secondary users”. The answer to this question depends on
various factors like the degree of cooperation between the primary and secondary users, the
reliability of the spectrum sensing and sharing scheme, the capabilities of the secondary and
primary networks etc. This paper focuses on the QoS provisioning of the cognitive radio. Also the
integration of real time or practical learning capabilities is also a challenge in cognitive radios.
Investigation of applications of lesser-known soft computing techniques can result in several
benefits as it enables making decisions through predictions and also when accurate real time
information is not available.
In this work, we focus on two methods. Firstly the Spectrum Sensing, in which the reliability of
the spectrum sensing technique using fuzzy logic is studied and analyzed for QoS provisioning.
Secondly, Spectrum Management, to control the QoS of the cognitive radio system by
optimization of the system using genetic algorithm is studied and evaluated. We can combine the
two methods to collectively address it as Spectrum Optimization.
2. GENETIC ALGORITHM
The concept of genetic algorithm lies on the fact that the measure of success of an individual is its
fitness i.e. “survival of fittest”. It states that the best combination of genes and their resulting
chromosomes yields the strongest individual, which will survive the longest. The first step of GA
requires the solutions to be encoded into chromosomes. In case of cognitive radio scenario, the
configurable radio parameters like transmit power, modulation, coding rate, symbol rate, packet
size etc. represent genes of a chromosome. We have taken such parameters for the chromosome
construction; transmit power (P), bandwidth (BW) and symbol rate (Rs), modulation, bit rate (Rb)
and one environmental parameter as noise power (N). These six genes will define the structure of
the chromosome. We have considered the values of the parameters as discussed [2], for the
implementation of cognitive radio as per IEEE 802.22 standard.
The structure of chromosome comprises of the above-mentioned six parameters or genes. The
parameters or genes are concatenated to form a string (chromosome). Table 1 gives the
summarized values of the order of the genes, the ranges of operation and the binary bits required
to encode the corresponding integer values.

3
Table 1 Summarized Values of the Chromosome Structure
Order 1 2 3 4 5 6
Gene Power Bandwidth Symbol
Rate
Bit
Rate
Modulation Noise
Power
Ranges -
30dBm
to
30dBm
40 to 840
MHz
1Mbps
to
8Mbps
0 to 2
Mbps
- -20
dBm -
-15
dBm
Step
Size
1 dBm 8 MHz 1 Mbps 125
Kbps
- 1
dBm
Integer
Value
Range
0 – 59 0 – 99 0 – 7 0 –
15
0 – 3 0 – 5
Binary
Bits
Required
6 7 3 4 2 3
The six parameters specified as genes in chromosome structure need 25 bits for its construction.
The composition of this bit string is important because the mutation operation performs at bit
level. Let us assume the parameters be x1, x2, x3, x4, x5 and x6 corresponding to the power,
bandwidth, symbol rate, bit rate, modulation and noise power respectively. Fitness functions for
each parameter are generated by the following formula initiated in [3] is given as:
. d
i i i d d
i i i id
i
w x x
f if x x x
x
 −
 = − <
   … (1)
where xi
d
is the required QoS parameter. wi is the weight subject to
6
1
1,i
i
w
=
=∑
where i =1, 2, …,6.
Overall fitness function ftotal is the cumulative sum of individual fitness functions of the
parameters given as:
6
1
total i
i
f f
=
= ∑
… (2)
The overall fitness value of the chromosomes in percentage is
Total fitness value (%) =
6
1
100 1 i
i
f
=
 
− 
 
∑
… (3)
Ideally each wi should be 16.7 %, that means each gene will have the same weight. But in
practical scenario, the weighing factor w can vary according to QoS specifications. Selection - We
compare the total fitness of the chromosomes using equation 3 and the best solutions from pool of
chromosomes are selected using ‘Roulette wheel selection’. It states that the probability of
selection for each individual (chromosome) is given by:
1
i
i n
i
i
f
p
f
=
=
∑
… (4)

4
pi is the probability of selection of individual chromosome. fi is the fitness of the individual gene i.
n is the number of chromosome in the population. The individuals (chromosomes) having the
maximum probability value are transferred to the next generation. Subsequently, with each
generation the fittest chromosomes are selected and transferred to next generation and made new
population until the termination criteria is reached
Table 2 Performance Objective Functions and Objective Functions
S.No. Objective
Function
Equation Remarks
1 Power
Consumption(Pc)
sPc PR K= [4] P=Transmit Power
(dBm);
Rs= Symbol
Rate(kbit/s);
K=Signal Space
dimensions(e.g. For
PSK, K=1, for QAM,
K=2 );
2 Spectral
Efficiency(ηB) B
bR
BW
 =
Rb=bit rate (kbit/s);
BW= Bandwidth;
3 Throughput (T)
max.
iM
T
Ni M
=
Mi=Modulation index
on channel i; Ni =
Number of channels;
Mmax=Max
Modulation Index
4 Signal to Noise
Interference
Ratio (SINR)
i
i
P
SINR
N I
=
+ ∑
,
where N = kTBW
P=Transmit
Power(dBm);
I = Interference ;
k = Bolzmann
constant =
1.38065x10-23
J/K;
T=Temperature;
BW =Bandwidth.
Table 2 shows the performance objectives to be achieved to lead a system to an optimal state. We
have specifies power consumption which has to be minimized, spectral efficiency which has to be
maximized, throughput has to be maximized and signal to noise ratio to be maximized.
3. CONVENTIONAL GENETIC ALGORITHM
The QoS requirements of different parameters are evaluated for different channel bandwidths.
This algorithm will have the flexibility to allow tradeoff between power consumption and spectral
efficiency according to the specific QoS requirements. The algorithm is tested over a number of
different initial GA settings and different adjustable parameters to compare numerical
optimization performance.
Elitism – In this dissertation, we have used one more method of selection called elitism. With the
process of elitism, some of the best chromosomes are selected from the current population and are
transferred directly to the next generation before the crossover and mutation operations. We have
set elitism rate of 2%. Crossover – Two chromosomes are selected from the current generation of
population, arbitrarily and multi – point crossover operation is applied giving new pair of

5
chromosomes. We have taken crossover rate of 90 %. This means that 90 out of 100
chromosomes will undergo crossover operation. Mutation – In this operation some randomly
selected bits are made to flip and fit into any of any of the six genes of the chromosome. This
operation will provide new solution space to the algorithm. The mutation rate we have taken is
5%. This rate can be increased to increase the fitness of the chromosomes. Termination criteria –
The steps above selection, crossover and mutation will repeat until termination criteria (best
solution) is reached. We have used two termination criteria. Firstly, the number of generations is
set to 1000 generations; secondly the convergence of population is set at .95. This means that as
the generation progresses and reaches 95% (means 95% of the chromosomes share the same
gene), the iterations are stopped. Here the average fitness approaches the best individual and the
best solution is selected.
Drawback of conventional genetic algorithm
While applying genetic algorithm to solve optimization problem, we come across a phenomenon
is called premature convergence [5]. When premature convergence occurs, the algorithm is
converged too fast to a solution, which causes it to be trapped in local minima. The main cause of
premature convergence is the lack of diversity because of the ineffectiveness or crossover and
mutation operators. Because of premature convergence, the evolutionary algorithm cannot
progress and some important genetic solutions are lost. To improve the performance of genetic
algorithm we address this issue in our next section. We propose adaptive genetic algorithm with
adapting operators and parameters. This is done by controlling the operators and parameters in
such a way that they will change the values if the population is not producing enough fit
individuals. This in turn will affect diversity and hence the performance of the algorithm will
increase. Moreover, it will increase the adaptability of the cognitive radio that is its one of the
main characteristic. The performance of the genetic algorithm is calculated by the average fitness
convergence values. The higher the fitness score, the fitter the individuals in that generation and
therefore the best solution can be achieved.

6
Table 3 Pseudo Code for a Conventional Genetic Algorithm
4. PROPOSED ADAPTIVE GENETIC ALGORITHM
The crossover and mutation operations are very important in the genetic algorithm for
optimization in cognitive radio. Population adaptation was proposed by Newman et al. [3] in
which the results from previous evolutions were utilized by seeding percentage of the initial
generation with chromosomes from final generation of the last cognition cycle. This method
improved the convergence rate of the algorithm to quite an extent. Subsequently, in [6] parameter
adaptation in multicarrier environment was proposed by cognitive radio scenario. It used binary
ant colony optimization method to optimize the cognitive radio parameters given a set of
objectives. The automation of the parameter and operator settings in genetic algorithm for
optimization in cognitive radio still remains an unexplored area so far. We have focused our
research on this issue. Therefore, we propose a generic scheme for adaptive crossover and
mutation rates.
Inputs
α – Size of population
γ – Rate of mutation
δ – Number of Generations Output X – Solution
λ – Rate of crossover
Begin:
// Initialization
1. Sense the environment for inputs;
2. Convert the inputs into chromosomes by encoding them;
3. Generate α feasible random solutions;
4. Save them in population Pop;
// Loop until the termination criteria is reached
5. for i = 1 to δ do;
// Selection Based on elitism and fitness function
6. ne = α.β;
7. Selection of best ne solutions in Pop and save them in Pop1;
//Crossover
8. crossover nc = (α-ne).λ;
9. for j = 1 to nc do;
10. Select two solutions randomly from Pop, XA and XB;
11. Generate XC and XD by two point crossover to XA and XB;
12. Save XC and XD to Pop2;
13. end for
//Mutation
14. for j = 1 to nc do;
15. Select a solution Xj from Pop2;
16. Mutate each bit of Xj under the rate γ and generate a new solution Xj’;
17. If Xj’ is unfeasible ;
18. Update Xj’ with a new feasible solution by repairing Xj’;
19. end if
20. update Xj’ with Xj’ in Pop2;
21. end for
// Updation of Population
22. update Pop = Pop1 + Pop2;
23. end for
// Best Solution
24. Return the best solution X in Pop;

7
In conventional genetic algorithm, the genetic operators such as crossover and mutation work
with a constant probability. However if we have different crossover and mutation rates, the
algorithm traverses different search directions in the search space thus improving the
performance. The essential design criterion of a cognitive radio is to adapt itself to the
surrounding changing environment. Due to this dynamic changing scenario, the proposed method
depends on maintaining an acceptable level of productivity throughout the process of evolution
and thus it is essential that the designed algorithm adapt itself to the appropriate crossover and
mutation rates, and thus optimizing out cognitive radio scenario.
Let the probability of crossover be pc and probability of mutation be pm. We use the progress
value concept as in [7] for calculating the progress (adjustment rates) for the probabilities. Let XA
and XB be two offsprings after crossover operation. Let FSum-S be the fitness sum of two offsprings
and FSum_P denotes the sum of the parent individuals.
Value CP = FSum_S + FSum_P … (5)
The generation undergoes nc crossover operations. Average crossover progress is given by CP :
1
c
CP CP
n
= ∑
… (6)
CP measures the overall performance of crossover operator in a generation. Similarly, the
overall performance of mutation operator in a generation is calculated as follows:
MP = fnew + fold … (7)
where fnew is the fitness of new offspring and fold is the fitness of the old offspring after mutation.
1
m
MP MP
n
= ∑
… (8)
The crossover and mutation operator performances are calculated each generation and adjusted
using the average progress values. The adjusted values of operators are as follows:
pc = pc + µ1 if CP MP> … (9)
pc = pc - µ1 if CP MP< … (10)
and pm = pm + µ2 if CP MP> … (11)
pm = pm - µ2 if CP MP< … (12)
where µ1 and µ2 represent the amount of adjustments of pc and pm. Experimental results in [8]
gives the values of the adjustments as:
max
1 2
max min
0.01
avgf f
f f
 
−
= =
− if fmax> fmin … (13)

8
1 2 0.01 = = if fmax=favg … (14)
Table 2 Pseudo Code for Proposed Adaptive Genetic Algorithm
5. SIMULATION RESULTS
Our simulation is targeted to determine the convergence speed of the fitness function using
conventional genetic algorithm. We also analyze the fitness value the system converges to.
Weights are defined in Table 6, which are used to create four different search modes for the
cognitive radio. Mode 1 is the low power mode (for applications e.g. text messaging), Mode 2 is
the high efficiency mode (e.g. ECG data), Mode 3 is the high throughput mode (e.g. broadband
video download) and Mode 4 is specifies for high reliability (e.g. defense services).
Inputs
α – Size of Population
γ – Rate of Mutation
δ – Number of Iterations
λ – Rate of Crossover Output X – Solutions
Begin
// Initialization
1. Sense the environment for inputs;
2. Convert the inputs into chromosomes by encoding them;
3. Generate feasible random solution;
4. Save them in Pop;
// Loop1 until termination criteria is reached
5. for i = 1 to δ do;
// Selection based on fitness function
6. Selection of best individuals in Pop;
7. Save them in Pop1;
//Loop2 Start
// Crossover and Mutation
8. Select two parents from Pop1;
9. Perform Crossover operation;
10. Save the offspring values in Pop2;
10. Accumulate the crossover progress value CP;
11. Perform Mutation Operation;
12. Mutate each bit of Xj in Pop2 under the rate γ and generate a new solution Xj’;
13. Update the offspring value Xj’ in Pop2;
14. Accumulate the mutation progress value MP;
// Loop2 end
// Progress Rates and Adjustments
15. Calculate CP ;
16. Adjust crossover rate λ;
17.Calculate MP ;
18. Adjust Mutation rate γ;
// Updation of Population
19. Update Pop= Pop1+Pop2;
// End Loop1
// Best Solution
20. Return the best solution X in Pop;

9
Table 5 Genetic Algorithm Settings
S.
No.
Parameter Value
1. Population Size 50
2. Length of
Chromosomes
25 bits
3. Number of
Generations
1000
4. Crossover Rate 90%
5. Mutation Rate 5%
6 Elitism Rate 2%
We aimed at targeting the speed of convergence of the fitness function using CGA(Conventional
Genetic Algorithm) and AGA (Adaptive Genetic Algorithm). Figure 1 shows the comparison of
the fitness convergence for CGA and AGA in lower power mode (Mode 1), Figure 2 shows the
comparison of the fitness convergence scenario for CGA and AGA in high efficiency mode
(Mode 2). Figure 3 shows the comparison of the fitness convergence scenario for CGA and AGA
in high throughput mode (Mode 3). And Figure 4 shows the comparison of fitness convergence
scenarios for CGA and AGA in high reliability mode (Mode 4).
Table 6 Weight Values
Objective w1 w2 w3 w4
Mode 1 0.65 0.15 0.08 0.12
Mode 2 0.12 0.65 0.15 0.08
Mode 3 0.08 0.12 0.65 0.15
Mode 4 0.15 0.08 0.12 0.65

10
Figure 1 Comparison of Average Fitness Scores of CGA and AGA for Mode 1
(Low Power Mode)
(High Efficiency Mode)
(High Throughput Mode)

11
(High Reliability Mode)
The results clearly show that the AGA (Adaptive Genetic Algorithm) outperforms the CGA
(Conventional Genetic Algorithm) in all the four modes of operation. The fitness convergence
values for the two algorithms are compared in Table 7.
Table 7 Comparison of Fitness Convergence Values
Generation
Number
Mode 1 Mode 2 Mode 3 Mode 4
CGA AGA CGA AGA CGA AGA CGA AGA
100 .8526 .8967 .7236 .8765 .6833 .8345 .8491 .9345
500 .9033 .9287 .7625 .8914 .6912 .8582 .8782 .9647
1000 .9567 .9654 .7989 .9581 .7087 .8999 .9579 .9737
It was analyzed that the best solution with the fitness value of the gene converging to zero, but in
our result the converging point is 0.2016 as shown in Figure 5.
Figure 5 Overall Fitness Value Plot for AGA

12
Figure 6 Generation Vs. Parameter Value (%) for AGA
With the proposed algorithm, it is found that the best value of the parameter to obtain the required
QoS specification for the cognitive radios system is for the frequency as shown in Figure 6. The
percentage of the best total fitness values against the average fitness values for AGA is shown in
Figure 7. It shows that the proposed algorithm has 84 % of best fitness values, which is more than
the previous algorithms proposed by the researchers.
Figure 7 Plot of the Fitness Values in (%) for AGA

13
Table 8 Comparison for the transmission parameter values after 1000 iterations.
Objective Average Power
(Pt) (dBm)
Bit Rate (Rb)
(Mbps)
Modulation
index (Mi)
CGA AGA CGA AGA CGA AGA
Low Power
Mode
3.267 2.763 0.687 0.834 4.768 2.76
High
Efficiency
Mode
10.45 9.76 0.852 0.966 4.891 3.48
High
Throughput
Mode
12.678 10.456 1.783 1.968 8.873 7.12
High
Reliability
Mode
5.463 4.113 1.569 1.683 6.873 5.76
Finally, the values of the transmission parameters were calculated and analyzed for both the
algorithms as summarized in Table 8. It shows that the goal was achieved in every mode of
operation. After 1000 iterations, the cognitive radio was able to decrease the power to an average
of 2.763 dBm. For the second goal to increase the efficiency the power had to be increased but
the modulation index was reduced. For high throughput mode, the power is increase to increase
the bit rate for high throughput transmission purposes. And for the fourth mode to increase the
reliability the balanced approach is applied, the power is kept not too high or not too low for
ensuring the reliability of the transmission.
6. CONCLUSION
In this paper, we developed performance objective fitness functions representing the relationship
between the transmission and environmental parameters that were identified in the literature
survey. For the implementation to be simple, we have not considered all the parameters, which
contribute to the cognitive radios system but a set of parameters that contribute to good extent.
The parameters were encoded using binary encoded method into chromosomes. The chromosome
of 25-bit length string was used. We first carried out the simulation scenario on conventional
genetic algorithm (CGA). The pseudo code was developed for step-by-step implementation of the
genetic algorithm implementation. The same set of conditions were tested upon the proposed
adaptive genetic algorithm (ADA) in which the crossover and mutation rates were automated.
This automation was based upon the previous fitness values and previous crossover and mutation
rates. The progress _value function was initialized, which measures the overall performance of
the crossover and mutation operators in a generation. The required adjustments are made for the
crossover and mutation rates that are decided by the progress _value function. The results of the
proposed algorithm show that the optimal solutions give the values for individual genes that
accommodate the secondary user with a specific QoS requirement. But the tradeoff between the
parameters has to be managed as the behavior is not linear. The simulation results shows that the
proposed method gives better solution to the cognitive system with adapting capabilities and

14
converges faster than the conventional methods of optimization of cognitive radio parameters.
The QoS parameters optimized by the proposed algorithm have higher fitness values and also
have the capability to manage the tradeoff between multiple objectives more effectively. The
convergence plots illustrated the expected trends but the best convergence fitness value reaches
stability at 95% in case of the proposed adaptive genetic algorithm. The problem of premature
convergence was also solved as the proposed adaptive genetic algorithm adjusts to the suitable
crossover and mutation rates to reduce the effort of searching for appropriate crossover and
mutation rates.
REFERENCES
[1] Federal Communications Commission - First Report, and Order and Further Notice of Proposed
Rulemaking, “Unlicensed operation in the TV broadcast bands," FCC 06-156, Oct. 2006.
[2] V. Blaschke, T. Renk, F. K, Jondral, “A Cognitive Radio Receiver Supporting Wide-Band Sensing,”
IEEE Conference on Communications, ICC ’08, pp. 499-503, 2008.
[3] T. R. Newman, R. Rajbanshi, A. M. Wyglinski, J. B. Evans, and G. J. Minden, “Population adaptation
for genetic algorithm-based cognitive radios," in IEEE Proceedings of Cognitive Radio Oriented
Wireless Networks and Communications, Orlando, FL, August 2007.
[4] T. W. Rondeau, B. Le, D. Maldonado and C.W. Bostian, “Cognitive Radio formulation and
implementation” in Proceedings of the First International Conference on Cognitive Radio Oriented
Wireless Networks and Communications, (CROWNCOM’06), 2004.
[5] Y. Leung, Y. Gao and Z. Hu, “Degree of population diversity – a perspective on premature
convergence in genetic algorithm and its markov chain analysis,” IEEE Transactions on Neural
Networks, Vol. 8, pp. 1165 – 1176, 1997.
[6] M. Waheed and C. Cai, “Cognitive Radio Parameter Adaptation in Multicarrier Environment,” IEEE
Fifth International Conference on Wireless and Mobile Communications, pp. 391 – 395, 2009.
[7] T. P. Hong, H. S. Wang, W, Y. Lee, “Evolution of appropriate crossover and mutation operators in a
genetic process,” Applied Intelligence, Vol. 16, pp. 7 – 17, 2002.
[8] W.Y. Lin, W.Y. Lee, T.P. Hong, “Adapting Crossover and Mutation Rates in Genetic Algorithm,”
Journal of Information Science and Engineering, pp. 1- 17, 2003.
Authors
Maninder Jeet Kaur is working as a Research Scholar in Department of Computer
Science Engineering at Dr B R Ambedkar National Institute of Technology,
Jalandhar, India. She has completed her B.Tech in Electronics and Communication
Engineering from Punjab Technical University in 2005. She completed her M.Tech
in Computer Science Engineering from Punjab Agricultural University, Ludhiana,
Punjab, India in 2007. She has published and presented many papers in
International Journals and Conferences. She is a member of International
Association of Engineering (IAENG) and International Association of Engineering
and Scientists(IAEST). She was selected for Commonwealth Split Site Doctoral
Fellowship-2010 for doing research work at University of York, United Kingdom for a period of 1 year.
Her current research interests includes Cognitive Radio, Artificial Intelligence, Information
Communication etc.

15
Moin Uddin , presently Pro – Vice Chancellor of Delhi Technological University and
Forner Director Dr B R Ambedkar National Institute of Technology, Jalandhar
(India). He obtained his B.Sc. Engineering and M.Sc. Engineering (Electrical) from
AMU, Aligarh in 1972 and 1978 respectively. He obtained hid Ph. D degree from
University of Roorkee, Roorkee in 1994. Before joining NIT, Jalandhar, he has
worked as Head Electrical Engineering Department and Dean Faculty of Engineering
and Technology at Jamia Millia Islamia (Central University) New Delhi. He
supervised 14 Ph. D thesis and more than 30 M.Tech dissertations. He has published
more than 40 research papers in reputed journals and conferences. Prof. Moin Uddin
holds membership of many professional bodies. He is a Senior Member of IEEE.
Harsh K. Verma received his PhD degree in Computer Science and Engineering from
Punjab Technical University, Jalandhar and Master’s degree from Birla Institute of
Technology, Pilani. He is presently working as Associate Professor in the Department
of Computer Science and Engineering at Dr B R Ambedkar National Institute of
Technology, Jalandhar, Punjab, India. He has published more than 20 research papers
in various Journals and Conferences of International repute. His teaching and research
activities include Scientific Computing, Information Security, Soft Computing and
Software Engineering.

OPTIMIZATION OF QOS PARAMETERS IN COGNITIVE RADIO USING ADAPTIVE GENETIC ALGORITHM

Recommended

More Related Content

What's hot (19)

Similar to OPTIMIZATION OF QOS PARAMETERS IN COGNITIVE RADIO USING ADAPTIVE GENETIC ALGORITHM (20)

More from ijngnjournal (20)

Recently uploaded (20)

OPTIMIZATION OF QOS PARAMETERS IN COGNITIVE RADIO USING ADAPTIVE GENETIC ALGORITHM