Abstract
Wireless Sensor Networks (WSN) are mainly utilized for time sensitive applications such as forest fire detection systems and health monitoring systems. Sensor nodes are operated on low power and limited computation process. It is essential to develop the solution for planning the topological area. Multiple sinks are located in the network and reduce the number of hops between the sensors and its sinks. We propose an efficient technique based on Bacteria Foraging Algorithm to identify the best optimal locations of sinks. The experimental results show that average end to end delay is minimized and average energy consumption of sensor nodes are reduced.
1. Introduction
Wireless Sensor Networks (WSN) are a collection of small sensor nodes with limited energy constraints that collect the information from the environment and forward it to the sink node through the intermediate nodes. Most of the applications of WSN are time sensitive such as forest fire detection, health monitoring, intrusion detection, security, surveillance, military applications and habitat monitoring. The sensor nodes have been restricted in the process of computation and communication.
Sensor nodes are placed in a random fashion, and their locations are not known exactly. Planning and controlling of the topology are the most important characteristics to handle the resources for processing in the WSN. Multiple sink placement is one of the methods to control the topological area. The station where the information is collected for the processing also called the base station or sink. A sink node has higher energy than the other sensor nodes. The number of optimal sinks is included in the network to manage the resources effectively. The optimal sink consumes the minimum energy consumption and improves the network lifetime. The placement of the multiple sinks is shown in Figure 1. Multiple sink placement problems can be solved by using the biologically inspired mechanism.
Multiple Sink Placement Problem.
The biologically inspired mechanism [17] is mainly used to design the sensor networks with scalability and robustness. The complex system behavior is adapted to the new environment by using the simple rules for solving the problem. The group behavior of the nature spices or insects find the solution for the distributed problem without external guidance. These methods are mainly used to find the solution for routing and clustering problems in the network.
Bacteria Foraging Algorithm (BFA) is a well-known approach in the biologically inspired mechanism to find the optimal number of sinks from the list of candidate positions in the sensor nodes. The best sink locations are utilized to minimize the average end to end delay and reduce average energy consumption.
The rest of the paper is structured as follows. Section 2 describes the related works for multiple sink placement and biologically inspired algorithms. Section 3 discusses the method to locate optimal multiple sinks by using BFA. Section 4 presents the performance analysis and results. Section 5 gives the conclusion.
2. Related Works
Kim et al. [7] have proposed to find the best location of multiple sinks and paths for the sensor networks. Oyman and Ersoy [11] have developed the multiple sink network design problem that consumes minimum operational time for the sensor networks. Schmitt and Poe [21] have introduced the genetic algorithm sink placement method used to locate the optimal solutions for the sink placement in the sensor networks. Poe and Schmitt [16] have developed the genetic algorithm utilized to reduce the delay for the multiple sinks in the WSN. Safa et al. [18] have proposed the discrete particle swarm optimization (DPSO) used to optimize the number of sinks and their position.
Das et al. [4] have described the partition techniques used to find the candidate location by using minimum hop count and centroid value of the nodes in the sensor networks. Sitanayah et al. [23] have proposed the methods for placing the multi sinks and relay placement for minimizing the cost value in the deployment process. Ciciriello et al. [3] have discussed the routing techniques used to identify the optimal solution for transferring the information effectively from multiple sources to sinks.
Ren and Meng [17] have proposed the bio inspired methods such as ant colony optimization and particle swarm optimization and designed the paradigm for the sensor networks. Iyengar et al. [5] have developed biological methods and utilized the resources effectively for designing of the adaptive routing schemes in the sensor networks. Selvakennedy et al. [22] have discussed the T-ANT (Ant colony Odynamic clustering protocol) used to identify the optimal number of cluster heads from the group of clusters that maintain the optimal performance for the entire network.
Kulkarni and Venayagamoorthy [8] have introduced the bio inspired techniques such as particle swarm optimization and BFA utilized for deploying the sensor nodes based on the segmentation of the images. Kim and Cho [6] have developed the hybrid method of genetic algorithm and BFA applied for tuning the proportional integral derivative (PID) controller.
Panigrahi and Pandi [13] have proposed the hybrid BFA with Nelder-Mead method utilized to identify the search space for the economic dispatch problem. Tang et al. [24] have developed the dynamic BFA utilized to identify the solution for optimal power flow in dynamic environment. Tripathy and Mishra [25] have described the BFA applied to optimize the power loss and voltage stability in the unified power flow controller. Pitchaimanickam and Radhakrishnan [15] have proposed to generate the optimal clusters from the group of nodes and identified the suitable cluster head from each cluster.
Panda and Triapthy [12] have discussed the modified BFA and solved the scheduling problem by generating the optimal schedule of the wind energy conversion system. Mishra [10] has developed the BFA for generating the optimal coefficients in the active power filter. Abd-Elazim and Ali [1] have proposed the BFA employed for controlling the optimal parameter in the static var compensator design of multi-machine power system.
Chen et al. [2] have presented the BFA which describes the chemotactic process for the communication between the cells to find the solution of radio frequency identification (RFID) network planning. Sanyal et al. [19] have discussed the adaptive bacterial foraging algorithm for optimizing the fuzzy entropy in the gray image segmentations. Sathya and Kayalvizhi [20] have proposed the modified bacterial foraging algorithm for performing the multi-level thresholding in the histogram image segmentation. Verma et al. [26] have presented the probabilistic derivative method from ant system which is used for the BFA to find the edge detection problem.
3. Proposed Method
3.1 Multiple Sink Placement Problem Using Bacteria Foraging Algorithm
In this section, the BFA is proposed to find optimal solution for the multiple sink placement. The main idea is to reduce the average end to end delay and average energy consumption. The region of indifference is described as the sink can be positioned anywhere in the sensor network without modifying the topology. This method is applied to collect the group of candidate locations. Each region can be represented by one candidate location, and the total number of regions NR is evaluated by the following equation.
where n and Nni are the number of sensor nodes and neighboring nodes. Figure 2 shows the region of indifference. From this figure the sensor node “A” has two neighbors as “B” and “C”. The number of regions for the sensor node “A” is calculated by using the equation (1) and obtaining the result 2∗2=4. The sensor node “A” will not be considered for other regions of the sensor nodes. The sensor node “B” has one neighbor which is “C”.
Region of Indifference.
The number of regions for sensor node “B” is 4+2∗1=6. Sensor node “C” has no neighbors, and the number of regions for “C” is 6+2∗0+1=7. Candidate locations are identified by using the discretization process. Figure 3 shows the grid points placed over the sensor field. The distance between the grid and sensor node is calculated. The resultant value is compared to the transmission range of the sensor node, and the candidate location is less than the radius. If the conditions are satisfied, the current node is included in the candidate location.
Modify the Grid Size.
At the end of the process, all the grid points are joined with the similar neighborhoods and one node is chosen from the particular region. The number of accepted grid points is evaluated with the value of NR. If the value is less than NR, then the grid size is increased. This method is repeated for the total number of regions [16]. This can be shown in Algorithm 1.
Candidate Location
| Read the positions of the sensor nodes. |
| Compute the total number of regions NR. |
| Modify the position of the grid size. |
| Identify the candidate location to every region. |
| For all the node n |
| If candidate position from the node position is less than transmission range and candidate position from the origin is less than radius then Increment the candidate location. |
3.2. Bacteria Foraging Algorithm
BFA algorithm was developed by Passino [14] and Liu and Passino [9] to find the solution for optimization. In the foraging process, animals find and get the food resources for maximizing the energy intake over the unit time. BFA algorithm, motivated by the foraging activities of Escherichia coli bacteria, provides the methods of searching and handling the food. This algorithm consists of chemotaxis, swarming, reproduction and elimination and dispersal processes.
Chemotaxis: This is the movement of the bacteria generated by using the group of flagella in the foraging process. This process is performed with the two different modes such as tumble and swim. In the swimming process, bacteria go to the desired location for searching the food resource. The bacteria move to the different location for the tumbling process. These two types of process are executed alternatively for the whole life of the bacteria.
Swarming: The bacteria reaches the best location for their food searching process. It sends the attract signal to the nearest bacteria and calculates the relative distance between the bacteria which is added to the cost value. Bacteria with minimum cost value is chosen from the group of bacteria and merged together to form the density of the bacteria.
Reproduction: In this process, the highest value of the healthy bacteria are split into two groups, and the lowest value of the bacteria goes to the dead state. The new set of bacteria is maintained in the original location with the swarm population.
Elimination and dispersal: The immediate or slow changes in the location can execute the elimination or dispersal process. The lowest healthy value of the bacteria can be removed from the location. The highest healthy value of the bacteria is split into two new bacteria and moved into the new location.
Multiple sink placement problem using Bacteria Foraging Algorithm
| Begin |
| Step 1: Assign the location of the sink from the candidate location list. |
| Step 2: For each elimination dispersal “l” |
| Step 3: For each reproduction “k” |
| Step 4: For each chemotaxis “j” |
| a) For each ith bacterium 1 to S |
| b) Find the cost of candidate position J(i, j, k, l) and saved into Jlast. |
| c) Tumble: |
| To produce the random vector Δ(i). |
| d) Move: |
| |
| e) Swim: |
| for (m=0; m<Ns; m++) |
| if J(i, j+1, k, l)<Jlast then |
| Jlast=J(i, j+1, k, l) |
| end if |
| |
| else |
| m=Ns |
| end for |
| Calculate the cost value of new candidate position J(i, j+1, k, l). |
| Evaluate the average end to end delay for the updated candidate position. |
| The steps (a) to (e) are repeated for the chemotaxis process. |
| Step 5: Reproduction: |
| Candidate positions are arranged in the ascending order of Jhealth. The most healthy bacteria are split into two copies. These new bacteria are located in the same location. |
| If k<Nre |
| go to step 3 |
| else |
| go to step 6 |
| end if |
| Step 6: Elimination and dispersal: |
| for each bacterium 1 to S |
| Generate the random number “rn”. |
| If rn≤Ped then |
| bacteria move towards the new candidate location. |
| else |
| bacteria maintain the same candidate location. |
| end if |
| end for |
| Step 7: If l<Ned then |
| go to step 2 |
| else |
| stop |
| end if |
| Step 8: Find the best optimal locations of the multiple sinks and to calculate the average end to end delay. |
| End. |
Movement of the bacteria is achieved by tumbling or swimming. In the tumbling process, a random vector is generated and the bacteria are moved into different directions for searching the better position of the candidate. In the swimming process, the bacteria are moved into different directions of the candidate position. If the bacteria choose tumbling or swimming process that helps to reach the better candidate position. Once it finds the position, it sends the signal to other candidate positions to calculate the relative distance for the candidate position.
If the cost value of the new candidate position is compared to the old candidate position, then the position and cost value of the new candidate is updated, and the average end to end delay for the new candidate position is calculated. The swimming process is repeated for the maximum value of swim length. The better candidate positions are arranged in ascending order by using the health values. The bacteria with minimum health value expires, and the bacteria with maximum health value is divided into two copies and located in similar position. These steps are executed until it reaches the maximum value of reproduction. If the value of the random number is less than the random probability value, then the bacterium shifts toward the new candidate position; otherwise, it stays in the same candidate position. This process can be repeated for all remaining candidate positions and to identify the optimized sink locations from the candidate list. Figure 4 represents the flow chart for solving multi sink placement problem using BFA.
Flowchart for Optimal Multi Sink Placement Using BFA.
4. Performance analysis and Results
4.1. Simulation Model
Wireless sensor nodes are arranged in random distribution with the topography of 100×100 m2. The number of sinks is taken from two to eight sinks. At the beginning stage, each sensor node has equal energy level. The sensor and the sink nodes are immobile. The sensor nodes are transmitted over the range of 16 m with the data rate of 19.2 kbps. Ad hoc On-Demand Distance Vector (AODV) protocol is applied for transmitting the data to the nearest sensor nodes.
Experiments are implemented by using Network Simulator-2, and evaluating the performance of BFA is compared to the DPSO. Experiments are executed for the different random topologies, and the error graph is plotted with 95% confidence interval. Table 1 represents the parameters applied for conducting experiments. Tables 2 and 3 show the parameters utilized for BFA and DPSO.
Simulation Parameters.
| Parameters | Value |
|---|---|
| Number of sensor nodes | 20, 40, 60, 80, 100 |
| Number of sink nodes | 2–8 |
| Size of the topology | 100×100 m2 |
| Protocol | AODV |
| Transmission range | 16 m |
| Simulation time | 600 s |
| Data rate | 19.2 kbps |
BFA Parameters.
| Parameters | Value |
|---|---|
| Number of bacteria | 100 |
| Swim length of bacteria | 4 |
| Chemotactic iterations | 100 |
| Reproduction iterations | 4 |
| Elimination and dispersal iterations | 2 |
| Probability value of Ped | 0.25 |
| Run length vector | 0.05 |
DPSO Parameters.
| Parameters | Value |
|---|---|
| Size of the population | 40 |
| No. of generations | 25 |
| Cognitive probability value | 0.5 |
| Social probability value | 0.5 |
| Inertia weight | 0.9 |
4.2. Performance Metrics
The following metrics are utilized to measure the performance of the BFA and DPSO algorithms.
End to end delay is defined as the time consumed for the packet transmission from the source node to multiple sinks.
Energy consumption is defined as the amount of energy used for the execution process.
Figure 5 demonstrates the number of sinks versus average end to end delay. The number of sinks is restricted to two sinks. Increasing the number of sinks also reduces the average end to end delay. BFA method quickly identifies the optimal solution for multiple sink locations that decreases the average end to end delay. BFA technique clearly shows how to minimize the average end to end delay compared to DPSO.
Number of Sinks vs. Average End to End Delay.
From the experiment, we evaluate the average end to end delay for BFA and DPSO algorithms with random topologies by enhancing the network size from 20 to 100 nodes. BFA performs better than DPSO. Figure 6 shows that BFA method quickly identifies the optimal sink locations and reduces the average end to end delay of the sensor networks.
Sensor Nodes vs. Average End to End Delay.
Figure 7 illustrates the average energy consumed for the sensor nodes. Energy is efficiently used for the execution process of the sensor nodes that also prolongs the network lifetime. BFA approach consumes the minimum amount of average energy compared to the DPSO. BFA technique quickly finds the best sink positions of the network. It is evident that BFA method minimizes the average energy consumption which extends the network life time to a period.
Sensor Nodes vs. Average Energy Consumption.
5. Conclusion
This paper presents the BFA to find the optimal positions in the multiple sink placement problem. We implemented the BFA and DPSO algorithms for random topologies with different number of optimal sink positions. BFA quickly identifies the optimal locations for the sinks that reduce the average end to end delay and energy consumption of the network. The experimental results illustrate that BFA performs better than DPSO algorithm. We extend this work by using the hybrid approach of BFA with particle swarm optimization to identify the optimal sink locations in the multiple sink placement problem.
Acknowledgment
The authors would like to thank Kalasalingam University for supporting this work.
Bibliography
[1] S. M. Abd-Elazim and E. S. Ali, Bacteria foraging optimization algorithm based SVC damping controller design for power system stability enhancement, Electr. Power Energy Syst.43 (2012), 933–940.10.1016/j.ijepes.2012.06.048Search in Google Scholar
[2] H. Chen, Y. Zhu and K. Hu, Multi-colony bacteria foraging optimization with cell-to-cell communication for RFID network planning, Appl. Soft Comput.10 (2010), 539–547.10.1016/j.asoc.2009.08.023Search in Google Scholar
[3] P. Ciciriello, L. Mottola and G. P. Picco, Efficient routing from multiple sources to multiple sinks in wireless sensor networks, 4th European Conference on Wireless Sensor Networks, pp. 34–50, 2007.10.1007/978-3-540-69830-2_3Search in Google Scholar
[4] D. Das, Z. Rehena, S. Roy and N. Mukherjee, Multiple sink placement strategies in wireless sensor networks, IEEE Int. Conf. Commun. Syst. Netw., pp. 1–7, 2013.10.1109/COMSNETS.2013.6465578Search in Google Scholar
[5] S. S. Iyengar, H.-C. Wu, N. Balakrishnan and S. Y. Chang, Biologically inspired cooperative routing for wireless mobile sensor networks, IEEE Syst. J.1 (2007), 29–37.10.1109/JSYST.2007.903101Search in Google Scholar
[6] D. H. Kim and J. H. Cho, A biologically inspired intelligent PID controller tunning for AVR systems, Int. J. Control Automat. Syst.4 (2006), 624–636.Search in Google Scholar
[7] H. Kim, Y. Seok, N. Choi, Y. Choi and T. Kwon, Optimal multi sink positioning and energy-efficient routing in wireless sensor networks, Lect. Notes Comput. Sci.3391 (2005), 264–274.10.1007/978-3-540-30582-8_28Search in Google Scholar
[8] R. V. Kulkarni and G. K. Venayagamoorthy, Bio-inspired algorithms for autonomous deployment and localization of sensor nodes, IEEE Trans. Syst. Man. Cybernetics Part C Appl. Rev.40 (2010), 663–675.10.1109/TSMCC.2010.2049649Search in Google Scholar
[9] Y. Liu and K. M. Passino, Biomimicry of social foraging bacteria for distributed optimization: models, principles, and emergent behaviors, J. Optimiz. Theory App.115 (2002), 603–628.10.1023/A:1021207331209Search in Google Scholar
[10] S. Mishra, Bacterial foraging technique based optimized active power filter for load compensation, IEEE Trans. Power Deliver.22 (2007), 457–465.10.1109/TPWRD.2006.876651Search in Google Scholar
[11] E. I. Oyman and C. Ersoy, Multiple sink network design problem in large scale wireless networks, IEEE Int. Conf. Commun.6 (2004), 3663–3667.10.1109/ICC.2004.1313226Search in Google Scholar
[12] A. Panda and M. Triapthy, Optimal power flow solution of wind integrated power system using modified bacteria foraging algorithm, Electr. Power Energy Syst.54 (2014), 306–314.10.1016/j.ijepes.2013.07.018Search in Google Scholar
[13] B. K. Panigrahi and V. R. Pandi, Bacterial foraging optimisation: Nelder-Mead hybrid algorithm for economic load dispatch, IET Generation Trans. Dist.2 (2008), 556–565.10.1049/iet-gtd:20070422Search in Google Scholar
[14] K. M. Passino, Biomimicry of bacterial foraging for distributed optimization and control, IEEE Control Syst. Mag.22 (2002), 52–67.10.1109/MCS.2002.1004010Search in Google Scholar
[15] B. Pitchaimanickam and S. Radhakrishnan, Bacteria foraging algorithm based clustering in wireless sensor networks, 2013 Fifth International Conference on Advanced Computing (ICoAC), Chennai, India, pp. 190–195, 2013.10.1109/ICoAC.2013.6921949Search in Google Scholar
[16] W. Y. Poe and J. B. Schmitt, Placing multiple sinks in time-sensitive wireless sensor networks using a genetic algorithm, 14th GI/ITG Conference on Measurement, Modeling, and Evaluation of Computer and Communication Systems, pp. 253–268, 2008.Search in Google Scholar
[17] H. Ren and M. Q.-H. Meng, Biologically inspired approaches for wireless sensor networks, IEEE Int. Conf. Mechatronics Automat, pp. 762–768, 2006.10.1109/ICMA.2006.257686Search in Google Scholar
[18] H. Safa, W. El-Hajj and H. Zoubian, A robust topology control solution for the sink placement problem in WSNs, J. Netw. Comput. Appl.39 (2014), 70–82.10.1016/j.jnca.2013.04.009Search in Google Scholar
[19] N. Sanyal, A. Chatterjee and S. Munshi, An adaptive bacterial foraging algorithm for fuzzy entropy based image segmentation, Expert Syst. Appl.38 (2011), 15489–15498.10.1016/j.eswa.2011.06.011Search in Google Scholar
[20] P. D. Sathya and R. Kayalvizhi, Modified bacterial foraging algorithm based multilevel thresholding for image segmentation, Eng. Appl. Artif. Intel.24 (2011), 595–615.10.1016/j.engappai.2010.12.001Search in Google Scholar
[21] J. B. Schmitt and W. Y. Poe, Minimizing the maximum delay in wireless sensor networks by intelligent sink placements, Technical Report 362/07, University of Kaiserslautern, 2007.Search in Google Scholar
[22] S. Selvakennedy, S. Sinnappan and Y. Shang, A biologically-inspired clustering protocol for wireless sensor networks, Elsevier Comput. Commun.30 (2007), 2786–2801.10.1016/j.comcom.2007.05.010Search in Google Scholar
[23] L. Sitanayah, K. N. Brown and C. J. Sreenan, Planning the deployment of multiple sinks and relays in wireless sensor networks, J. Heuristics Arch.21 (2015), 197–232.10.1007/s10732-014-9256-zSearch in Google Scholar
[24] W. J. Tang, Q. H. Wu and J. R. Saunders, Bacteria foraging algorithm for dynamic environments, 2006 IEEE Congress on Evolutionary Computation, Vancouver, BC, Canada, pp. 1324–1330, 2006.Search in Google Scholar
[25] M. Tripathy and S. Mishra, Bacteria foraging based solution to optimize both real power loss and voltage stability limit, IEEE Trans. Power Syst.22 (2007), 240–247.10.1109/TPWRS.2006.887968Search in Google Scholar
[26] O. P. Verma, M. Hanmandlu, P. Kumar, S. Chhabra and A. Jindal, A novel bacterial foraging technique for edge detection, Pattern Recogn. Lett.32 (2011), 1187–1196.10.1016/j.patrec.2011.03.008Search in Google Scholar
©2018 Walter de Gruyter GmbH, Berlin/Boston
This article is distributed under the terms of the Creative Commons Attribution Non-Commercial License, which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
