Optimization of Average Distance Based Self-Relocation Algorithm Using Augmented Lagrangian Method

David C. Wyld et al. (Eds) : ACITY, DPPR, VLSI, WiMNET, AIAA, CNDC - 2015
pp. 193–203, 2015. © CS & IT-CSCP 2015 DOI : 10.5121/csit.2015.51317
OPTIMIZATION OF AVERAGE DISTANCE
BASED SELF-RELOCATION ALGORITHM
USING AUGMENTED LAGRANGIAN
METHOD
Shivani Dadwal1
and T. S. Panag2
1
Deptt of Electronics and Communication Engineering, BBSBEC,
Fatehgarh Sahib, India
shivanidadwal@gmail.com
2
Deptt of Electronics and Communication Engineering, BBSBEC,
Fatehgarh Sahib, India
tripatjot.singh@bbsbec.ac.in
ABSTRACT
Mobile robots with sensors installed on them are used in wireless sensor networks to generate
information about the area. These mobile robotic sensors have to relocate themselves after
initial location in the field to gain maximum coverage The average distance based algorithm
for relocation process of mobile sensors does not require any GPS system for tracking the
robotic sensors, thus avoiding cost, but increasing energy consumption. Augmented Lagrangian
method is introduced in average distance based algorithm to reduce the extra energy
consumption by sensors in average distance based relocation process. This modified average
distance relocation scheme also improves the coverage area and the time taken by mobile
robotic sensors to come to their final positions.
KEYWORDS
Average distance based self-relocation, modified average distance self-relocation.
1. INTRODUCTION
A lot of work has been done on sensor network as it has a wide application in various fields. One
of its major application is in military area wherein surveillance of enemy areas has to be done.
Earlier static sensors were deployed for detecting targets in a particular area. This reduced the
coverage area of sensors. With advances in mobile robotics, now sensors can be carried by
robotic structures. All they need is a direction which is administered by an algorithm. Execution
of these algorithm leads to deployment of sensors in an area in a manner that maximum area is
covered and all the targets are well detected. Even civilian application of such mobile sensors
exists in property and homeland security [1, 2].

194 Computer Science & Information Technology (CS & IT)
For military purposes, the sensors are mostly dropped from air, which leads to a random
deployment of sensors. These sensors have to reposition themselves. This issue has been covered
in our paper. Optimization of coverage has been an active topic in sensor networks as along with
coverage there are many other issues like minimum energy consumption and maximum lifetime
which have to be kept in mind. Optimization of static sensor networks has been done using
Genetic Algorithms which helps in locating sensors to their best position for maximum coverage
and saves energy resulting in increased lifetime [3,4,5,6]. Other than this a few distributed and
centralized algorithms have been introduced in an effort to modify and improve sensor
networking [7]. The Genetic algorithm for optimization cannot be used for distributed algorithms
as they require a central operating system to control the sensor positioning. Moreover Genetic
algorithm can be used only where the area is well known.
Many algorithms have been developed for placing the sensors evenly in a field, out of which
potential field algorithm is one [8]. In this algorithm, a potential field is generated by the sensors
and they move accordingly. Another algorithm is the virtual force algorithm [9] in which the
sensors move as per the attractive and repulsive forces generated by the sensors depending on the
distances between them and attain their final positions. Some other methods namely, density
control method [10] and fluid model based method [11] have also been introduced. There is
another algorithm that repairs the coverage by finding the terminated sensors and moves these
sensors to uncovered areas [12]. These algorithms listed above are applicable for distributed
network where there is no central node and each sensor decides its own path. For this kind of
network, a GPS system is must in order to track down the position of sensors after and before
relocation. GPS is global positioning system that consumes a lot of power and is costly. Every
time the sensor move during the execution of their respective algorithms, they send information
to the GPS system which again consumes power.
Some methods were introduced to decrease the power consumption of sensor out of which one
was to reduce the sensing range of sensor [13]. This was done on non-mobile sensors. It saves
energy and increases the lifetime of sensors. The average distance based relocation process does
not use any GPS system and hence cuts the energy consumption and cost as well [14]. But since
there is no GPS used, the sensors consume extra energy to find their best final position which
also increases the number of iterations to find the final coverage. Hence, optimization of average
distance based self-relocation process is introduced to facilitate the sensors to find their final
positions in less time which would also reduce energy consumption. This is done using
augmented lagrangian optimization method [19]. In the next section, average distance based self-
relocation algorithm has been discussed.
2. AVERAGE DISTANCE BASED SELF RELOCATION PROCEESS
2.1. Assumptions
Following are a few assumptions that we consider for the algorithm:
• Each sensor has a sensing area in the form of a circle, having radius, r. The probability of
covering this area is 1.
• The area A, where the sensors are randomly deployed is known by the sensors approximately

Computer Science & Information Technology (CS & IT) 195
• If the sensors come within their sensing range Rc, the strength of the signals transmitted by
each sensor can be measured by the other.
• All sensors have certain range of communication and have a transmission power.
• Sensors have the ability to move as per the coordinates given after execution of algorithm.
• Sensors can detect obstacles in the field.
• Sensors that meet an obstacle in the field has its movement blocked and cannot communicate
with the rest of sensors.
2.2. Framework
The main aim of this algorithm is designing a distributed algorithm which has a self- relocation
capability to optimize the coverage area of field using less energy. It this algorithm, the distances
between the sensors have to be known in order to relocate the sensors. The sensors transmit
signal in the field once they are randomly deployed. These signals are intercepted by the sensors
that come within the reach of .sensor areas of other sensors. The received signal strength is
measured and corresponding distance is known.
Firstly, a “hello” signal is transmitted by the sensors which gives the signal strength to all the
sensors, near or far, lying in sensor range. The distance corresponding to this signal strength is
calculated by the sensors. Taking these distance information into account, the sensors move
towards or away to each other and relocate themselves. Also, any obstacle coming in way has to
be avoided by the sensors.
3. METHODOLOGY
3.1. Ideal Deployment
The ideal deployment is achieved when there are no spaces between the sensors. Such a condition
is possible when the distance between the sensors is √3r [15]. This is shown in Figure 1. Since it
is not possible to achieve an ideal condition, in this algorithm, we try to achieve a near to ideal
condition by placing sensors close. But as the field may have a deformity of obstacles, and
various other factors, such a condition is difficult to achieve.
3.2. Calculation of Threshold
The total number of sensors in the field and the sensing field area are used to calculate the
threshold distance dth and sensing radius that are near to ideal deployment. The threshold
distance dth decides the sensor movement.
Let the total area of the sensor field be A. As shown in Figure 1, assume that each sensor has an
effective area of coverage, E as in [14]. Let the total number of sensors deployed in the field be
N. Effective coverage of each sensor is given as:
‫ܧ‬ =
஺
ே
(1)

Also, E is a hexagonal area. Area of a hexagon is given as:
‫ܧ‬ =
ଷ√ଷ
ଶ
. ‫ݎ‬ଶ
(2)
The threshold distance and sensing radius are calculated by the following equations:
‫ݎ‬ = ට
ଶ
ଷ√ଷ
.
஺
ே
(3)
݀௧௛ = √3‫ݎ‬ (4)
The effective coverage, E should be larger than this value, as the sensors which lie close to
obstacles or to the edges will have less coverage. Hence, the threshold distance and sensor radius
are increased by 15%.
Figure 1. : Ideal coverage
3.3. Virtual Nodes
The algorithm considers that there exist virtual nodes at the boundary of the field. This is done to
make deployment easy for the sensors lying close to the edges. Such virtual nodes do not actually
exist. They are just used to avoid the sensors from getting closer to the edges. As shown in Figure
1, virtual nodes are considered along the boundary. No virtual nodes are required in the
optimization technique used in this work.
3.4. Movement Standards
The sensors relocate themselves by adjusting the distances between themselves. They either
move far or get closer to each other. No sensor has information about the direction of the other
sensor. The criteria in which the sensors move is described as below:
Standard 1: If there is at least one sensor in the communication range of sensor S having distance
less than 0.9dth, then the sensor will move away from other sensors.
Standard 2: If the standard 1 is not met and not more than 2 sensors lie at distance less than 1.1dth
from S, then the sensor needs to move closer to other sensors.
Standard 3: Sensor S need not move if the above two standards are not met.

A 10% margin is kept along these standards, so that the sensor is able to achieve a distance nearer
to dth from the rest of sensors.
3.5. Moving Distance
The movement of sensors is based on two standards, hence the moving distance is calculated by
the following equation:
݀௧௥௔௩௘௟ =
‫ە‬
‫۔‬
‫ۓ‬݀௧௛ −
ଵ
௠భ
∑ ݀௝
௠భ
௝ୀଵ ݂‫ݎ݋‬ ‫݀ݎܽ݀݊ܽݐݏ‬ 1
ଵ
௠మ
∑ ݀௜
௠మ
௜ୀଵ − ݀௧௛ ݂‫ݎ݋‬ ‫݀ݎܽ݀݊ܽݐݏ‬ 2
0 ݂‫ݎ݋‬ ‫݀ݎܽ݀݊ܽݐݏ‬ 3
(5)
In the above equation, dj and di are the distances from sensor S to other sensors. In the first
standard, the sensor moves only for the sensors closer than distance threshold. The total number
of sensors is given by m1. In the second standard, all the neighboring sensors to sensor S are taken
into account. The total number of sensors is m2.
The direction of movement of the sensor is chosen randomly as the sensor is unaware of the
neighbouring sensors. There can be a back and forth movement of sensors. To avoid this
direction control scheme is used. As the sensors move, the difference of direction of movement is
kept less than 90 degrees. Let the last direction of movement be α, then in the next movement
direction has to be in between α-90 degrees to α+90 degrees. When standard 1 is executed, the
sensors move away from each other and when standard 2 is executed, the sensors move nearer to
each other. As the sensors are moving to the direction chosen randomly, they check after moving
through a short distance if the required coverage is attained. If not, they come back to their
original positions.
4. SIMULATION AND ANALYSIS
For examining our results, we consider that the sensing field is a 100 by 100 grid structure. Each
grid is 1 meter apart from the other grid. Consider that the sensing range of the sensors used lies
within 18 to 25 meters. Hence, the maximum sensing radius is 25 meters. The range of
communication of sensors is kept almost double the maximum sensing radius i.e. 55 meters.
4.1. A Average Distance Based Self- relocation Algorithm Performance
Let 20 sensors be randomly deployed in the sensor field. The distance threshold and sensing
radius can be calculated from equations (3) and (4). A 15% increase should also be considered as
explained in Section 2.3.
The following equations give the sensing radius and threshold distance of 20 sensors:
‫ݎ‬ = 1.15ට
ଶ
ଷ√ଷ
.
஺
ே
= ට
ଶ
ଷ√ଷ
ଵ଴଴×ଵ଴଴
ଶ଴
= 15.9 ݉݁‫ݏݎ݁ݐ‬
݀௧௛ = √3‫ݎ‬ = √3 × 15.9 = 27.6 ݉݁‫ݏݎ݁ݐ‬

To analyze the results, three different conditions of initial sensor placement can be considered. In
the first case, the sensors are all placed in the center of the sensing field such that they cover 50
by 50 meters area. In the second condition, the sensors are deployed or scattered in the whole
sensing area. In the third condition, sensor are divided into 2 groups which are separately placed
in the sensing field. The coverage initially is calculated for all the conditions which come out to
be 55%, 50% and 61% approximately. The coverage can be calculated by the following
equations:
ܴ௖௢௩௘௥௔௚௘ =
஺೎೚ೡ೐ೝ೐೏
஺೟೚೟ೌ೗
(6)
Here,
ܴ௖௢௩௘௥௔௚௘ = ܿ‫݁݃ܽݎ݁ݒ݋‬ ‫݋݅ݐܽݎ‬
‫ܣ‬௖௢௩௘௥௘ௗ = ܽ‫ܽ݁ݎ‬ ܿ‫݀݁ݎ݁ݒ݋‬ ܾ‫ݕ‬ ‫ݏݎ݋ݏ݊݁ݏ‬ ܿ‫ݕ݈݁ݒ݅ݐܽݎ݁݌݋݋‬
‫ܣ‬௧௢௧௔௟ = ‫݃݊݅ݏ݊݁ݏ‬ ݂݈݅݁݀ ܽ‫ܽ݁ݎ‬
If we have to find out the coverage from the 100 by 100 grid structure, then the following
equation can be used:
ܴ௖௢௩௘௥௔௚௘ =
௡
ே
(7)
Here,
n= number of grid points covered by sensors
N= total number of grid points.
4.2. Coverage Analysis
The sensors in this algorithm do not have a fixed direction of movement. They move randomly
and check if the required conditions are met. So, each of the three conditions stated in the above
section are run 10000 times with a desired round number of 20. The execution results are shown
in the Figure 2.
In figure 2, the results of average distance based algorithm are compared with virtual force
algorithm. The virtual force algorithm is executed by taking its parameters into account as given
in [16]. In all the three conditions, the coverage increases as the number of rounds increase. An
average distance based self-relocation algorithm can achieve nearly 94% of coverage in the
sensing field after 20 rounds. For a virtual force algorithm, after 20 rounds 93% coverage is
achieved.
Thus, both the algorithm lead to almost same coverage of sensing field by the sensors. The major
difference is that the average distance based algorithm does not require a GPS hardware, reducing
the cost of the sensor network. It can hence be used in those areas where GPS cannot operate, like
in under water systems.

Figure 2: Simulation results of average coverage vs round number for average distance based self –
relocation algorithm
4.2. Energy Analysis
An energy analysis model has been discussed in [17]. This model has been used for mobile robots
as per which energy consumed by a robot to move 1 meter is equal to 9.34 Joules if it is moving
at a speed of 0.08 m/s constantly. The amount of energy used by robots to turn by 90 degrees is
2.35 Joules. Both travelling and turning of mobile robots is considered to be at a constant speed.
By doing so, we can plot a linear graph between energy consumed by mobile robot and coverage.
As per the above discussion, we can divide energy consumption into two following parts:
• Energy used while travelling: If the mobile robot keeps moving at a constant rate in one
direction, then the energy consumed for one single sensor node in a single round is given
as:
‫ܧ‬௧௥௔௩௘௟ = ݀௧௥௔௩௘௟ × 9.34(‫)ݏ݈݁ݑ݋ܬ‬
• Energy used in direction changing: As discussed in the relocation algorithm, the sensor
move randomly in some direction and checks its position by recalculating the signal
strength. After this, it decides whether to move back to original position or not. Hence,
during this process, the energy used by the sensor to turn in some direction is given as:
‫ܧ‬௧௨௥௡ = ቊ
൫‫ܣ‬ௗ௜௙௙/90൯ × 2.35(‫)ݏ݈݁ݑ݋ܬ‬ ‫݌݁݁ܭ‬ ݉‫݃݊ݒ݅݋‬
(360/90) × 2.35(‫)ݏ݈݁ݑ݋ܬ‬ ܶ‫݃݊݅݊ݎݑ‬ ܾܽܿ݇
In the above equation, Adiff is the difference of direction between the previous and later direction
of sensor movement. If the new position does not satisfy the coverage requirement of sensor, then
it gets back to position where it started by moving at 360 degrees, as shown in the above
equation.

To plot a graph, virtual force algorithm has also been considered. Its energy consumption is
calculated. In the virtual force algorithm, the above two energy consumption are added to the
energy used by the GPS system in locating the sensor nodes and in exchange of information
between sensor and GPS. The GPS chip uses 198 MW as in [18]. As the sensors are moving
constantly at 0.08m/s, the GPS consumes energy per meter given by following equation:
0.198 × (1/0.08) = 2.475 ‫ݎ݁ݐ݁݉/ݏ݈݁ݑ݋ܬ‬
As given in the Figure 3, virtual force algorithm will consume lesser energy as compared to
average distance based algorithm. The graph is plotted between average energy consumption and
average coverage for both the algorithms. The VFA uses lesser energy as it is GPS enabled,
making it easier for the sensor to redeploy themselves faster than the sensors in average distance
based algorithm. In VFA sensors take less time to redeploy as they know their positions
corresponding to other sensors and also know the locations of rest of the sensors in the field.
Figure 3: Simulation results for average coverage vs energy consumption
5. OPTIMIZATION USING AUGMENTED LANGRANGIAN METHOD
The basic drawback of average distance based algorithm is that it consumes more energy in
relocating the sensors as they move back and forth many times to come to an appropriate position
to get a good coverage.
Here, if an optimization technique is used to help sensors in their relocation process, they can
come to final position in lesser time consuming lesser energy. In the average distance based self-
relocation process, virtual nodes are considered at the boundary and outside the boundary as
shown in figure 1. Using these virtual nodes, the sensors make an idea of their boundary and the
area beyond which they are restricted. The sensors move back and forth in self –relocation
process, during which the sensors nearer to the boundary have to recalculate its positions to stay

within the boundary, which can be energy consuming. Thus, in the augmented lagrangian
method, the sensors are applied a penalty function. Using this penalty function, if the sensor skips
outside the boundary by certain distance dout, it is made to to come inside the boundary by the
same distance dout, as measured from the boundary. During this process, the sensor might overlap
or come into the boundary of another sensor, violating the threshold distance conditions as
discussed in section 2.3.
In the augmented lagrangian optimization method, an objective function has to be considered.
The minimum distance between the sensors should be dth , as in average distance based relocation
process. This minimum distance might change when penalty function is applied, so minimum
distance between sensor i.e. threshold distance dth is the objective function subject to constraints
in augmented lagrangian optimization method.
Now let the boundary of the sensor field be defined by following functions.
݃ଵ > 0, ݃ଶ ≥ 0, ݃ଷ < 0, … , ݃௡ > 0
Here, ݃ଵ, ݃ଶ, ݃ଷ, … , ݃௡ are the boundary equations for sensor field and n is the number of
boundaries.
In augmented lagrangian optimization method, new objective function is given as
݂ሾܺሿ − ߙ〈݀௧௛〉 − ߚ〈݃ଵሾܺሿ + ݃ଶሾܺሿ + ⋯ + ݃௡ሾܺሿ〉
ሾܺሿ is the co-ordinate of all the sensors obtained after first iteration of average distance based
relocation algorithm. The outline of the algorithm is as follows:
• The ሾܺሿ co-ordinates of all the sensors in the sensing field are obtained after the
execution of first iteration of average distance based relocation algorithm.
• The boundary constraint violations are checked i.e. it is checked if any sensor is lying
on the boundary or beyond the boundary. The constraints are checked n times which is
equal to the number of sensors in the field.
• If any constraint is violated, let us say for nth sensor,ሾ݃ଵሿ௜ ≮ 0, which means that
constraint ݃ଵ is violated by ith sensor and the sensor is moving beyond the boundary of
sensing field. Hence, the distance Di is calculated between ݃ଵ and ሾܺ, ܻሿ௜. X and Y are
the co-ordinated of ith sensor lying outside the boundary.
• To compensate for distance and to bring the sensor back inside the field, negative of
distance Di is added in g1 and the new co-ordinates of ith sensor are calculated.
• Now, all the co-ordinates serve as initial solution for average distance based algorithm
which is run again.
• The threshold distance are again checked in the self -relocation algorithm.
Hence, this method is a modified version of average distance based self-relocation algorithm
using the properties of augmented lagrangian optimization technique.

6. FUTURE WORK
The biggest advantage of the relocation scheme discussed is non- requirement of a GPS hardware
which cuts cost and increases the applicability of this relocation process where GPS cannot be
used. The drawback of more energy consumption is to some extent improved by the optimization
technique, yet more work can be done to further lessen the energy consumption of non GPS using
sensors.
REFERENCES
[1] V. Potdar, A. Sharif and E. Chang, “Wireless Sensor Networks: A Survey,” Advanced Information
Networking and Applications Workshops, Bradford, 26-29 May 2009, pp. 636-641.
[2] Ding, W. and Marchionini, G. 1997. A Study on Video Browsing Strategies. Technical Report.
University of Maryland at College Park.
[3] J. Chen and C. Li, “Coverage Optimization Based on Improved NSGA-II in Wireless Sensor
Network,” IEEE International Conference on Integration Technology (ICIT), Shenzhen, 20-24 March
2007, pp. 614-618.
[4] X. Wang, S. Wang and D. W. Bi, “Dynamic Sensor Nodes Selection Strategy for Wireless Sensor
Networks,” 7th International Symposium on Communications and Information Technologies
(ISCIT), Sydney, 16-19 October 2007, pp. 1137-1142.
[5] J. Weck,”Layout Optimization for a Wireless Sensor Network Using a Multi Objective Genetic
Algorithm,” IEEE 59th Vehicular Technology Conference, Milan, Vol.5, 2004, pp.2466-2470.
[6] L.C. Wei, C. W. Kang and J. H. Chen, “A Force-Driven Evolutionary Approach for Multi-Objective
3D Differentiated Sensor Network Deployment,” IEEE 6th International Conference on Mobile Ad-
hoc and Sensor Systems (MASS), Macau, 12-15 October 2009, pp. 983-988
[7] S. Dadwal, T.S. Panag, “Sensor Deployment Strategies in WSNs,” 3rd National Conference on
Recent Advances in Electronics and Communication Technologies, Ludhiana, 21-22 March 2013, pp
412-419.
[8] A. Howard, M. Mataric and G. Sukhatme, “Mobile Sensor Network Deployment Using Potential
Fields: A Distributed, Scalable Solution to the Area Coverage Problem,” The 6th International
Symposium on Distributed Autonomous Robotics System, Fukuoka, 25-27 June 2002, pp. 299-308.
[9] Yi. Zou and K. Chakrabarty, “Sensor deployment and Target Localization Based on Virtual Forces,”
IEEE Societies Twenty Second Annual Joint Conference of the IEEE Computer and Communications
(INFOCOM), San Fransisco, 1-3 April 2003, pp. 1293-1303
[10] M. R. Pac, A. M. Erkmen and I. Erkmen, “Scalable Self- Deployment of Mobile Sensor Networks: A
Fluid Dynamics Approach,” Proceedings of IEEE International Conference on Intelligent Robots and
Systems (RSJ), Beijing, 9-15 October 2006, pp. 1446-1451.
[11] R.-S. Chang and S.-H. Wang, “Self-Deployment by Density Control in Sensor Networks,” IEEE
Transactions on Vehicular Technology, Vol. 57, No. 3, 2008, pp. 1745-755.
[12] G. Wang, G. H. Cao, T. F. Porta and W. S. Zhang, “Sensor Relocation in Mobile Sensor Networks,”
IEEE Societies 24th Annual Joint Conference of the IEEE Computer and Communications, Miami,
2005, pp. 2302-2312.
[13] M. Cardei, J. Wu, M. M. Lu and M. O. Pervaiz, “Maximum Network Lifetime in Wireless Sensor
Networks with Adjustable Sensing Ranges, “ IEEE International Conference on Wireless and Mobile
Computing, Net- working and Communications, Montreal, 22-24 August 2005, pp. 438-445
[14] Y. Qu, S.V. Georgakopoulos, ”An Average Distance Based Self-Relocation and Self-Healing
Algorithm for Mobile Sensor Networks,” Journal of Wireless Sensor Network, Vol. 4 No.11, pp. 257-
263,2012.
[15] G. Wang, G. H. Cao and T. F. Porta, “Movement-Assisted Sensor Deployment,” IEEE Transactions
on Mobile Computing, Vol. 5, No. 6, 2006, pp. 640-652.

[16] W. H. Sheng, G. Tewolde and S. Ci, “Micro Mobile Ro- bots in Active Sensor Networks: Closing the
Loop,” Proceedings of IEEE International Conference on Intelligent Robots and Systems (RSJ),
Beijing, 9-15 October 2006, pp. 1440-1445.
[17] Y. G. Mei, Y.-H. Lu, Y. C. Hu and C. S. G. Lee, “Energy Efficient Motion Planning for Mobile
Robots,” Proceedings of IEEE International Conference on Robotics and Automation (ICRA), New
Orleans, 25 April-1 May 2004, pp. 4344-4349.
[18] M. K. Stojcev, M. R. Kosanovic and L. R. Golubovic, “Power Management and Energy Harvesting
Techniques for Wireless Sensor Nodes,” 9th International Conference on Telecommunication in
Modern Satellite, Cable, and Broadcasting Services, 7-9 October 2009, pp. 65-72.
[19] S. Singh, E. Singla, “Optimal design of robotic arms in constrained environment,” TEQIP-II
sponsored National conference on ‘Latest Developments in Materials, Manufacturing and Quality
Control’ MMQC, Bathinda , 18-19 March 2015.
AUTHORS
Shivani Dadwal had completed her B.Tech degree from RIEIT, Railmajra. Now she
is pursuing her M.Tech from BBSEBC, Fatehgarh Sahib.

Optimization of Average Distance Based Self-Relocation Algorithm Using Augmented Lagrangian Method

Recommended

More Related Content

What's hot (16)

Viewers also liked (20)

Similar to Optimization of Average Distance Based Self-Relocation Algorithm Using Augmented Lagrangian Method (20)

Recently uploaded (20)

Optimization of Average Distance Based Self-Relocation Algorithm Using Augmented Lagrangian Method