An optimization nodes layout in deployment WSN based on improved artificial bee colony

Chia sẻ: Lin Yanjun | Ngày: | Loại File: PDF | Số trang:10

Thêm vào BST

Báo xấu

27
lượt xem 4
download

Download Vui lòng tải xuống để xem tài liệu đầy đủ

This study suggests an optimal coverage model aimed at the optimal deployment of Wireless sensor networks (WSN) based on improving the artificial bee colony (ABC) algorithm. The distance between the sensor nodes is regulated reasonably by implementing quasi-gravitation and quasi-coulomb power. Also, with a low regional repetition rate, the ABC algorithm can achieve fast optimization. Besides, to minimize node energy consumption, the sensing radius of WSN nodes is dynamically modified.

Chủ đề:

Bình luận(0) Đăng nhập để gửi bình luận!

Lưu

Nội dung Text: An optimization nodes layout in deployment WSN based on improved artificial bee colony

An Optimization Nodes Layout in Deployment WSN Based on Improved Artificial Bee Colony Trong-The Nguyen2, Thi-Kien Dao1, Thi-Thanh-Tan Nguyen3, Truong-Giang Ngo4*, Duc-Tinh Pham5 1Fujian Provincial Key Laboratory of Big Data Mining and Applications, Fujian University of Technology, Fujian Province 350118, China 2Haiphong University of Management and Technology, Haiphong, Vietnam 3Information Technology Faculty, Electric Power University, Hanoi, Vietnam 4Faculty of Computer Science and Engineering, Thuyloi University, Hanoi, Vietnam 5Center of Information Technology, Hanoi University of Industry, Hanoi, Vietnam vnthe@hpu.edu.vn, jvnkien@gmail.com, *giangnt@tlu.edu.vn, tanntt@epu.edu.vn, tinhpd@haui.edu.vn Abstract. This study suggests an optimal coverage model aimed at the optimal deployment of Wireless sensor networks (WSN) based on improving the artificial bee colony (ABC) algorithm. The distance between the sensor nodes is regulated reasonably by implementing quasi-gravitation and quasi-coulomb power. Also, with a low regional repetition rate, the ABC algorithm can achieve fast optimi- zation. Besides, to minimize node energy consumption, the sensing radius of WSN nodes is dynamically modified. The simulation results show that the pro- posed algorithm performs better than the conventional particle swarm (PSO) and quantum-behaved particle swarm optimization (QPSO), ACO schemes in terms of network coverage rate, and convergence speed. Simultaneously, the algorithm has a particular benefit in reducing the energy consumption of nodes in the WSN. Keywords: Optimal deployment, Artificial bee colony, Quasi-physical strategy, Wireless sensor network 1 Introduction Wireless Sensor Network (WSN) has become a new hot and front research area[1]. Its subject combines sensor technology, embedded computing technology, distributed in- formation processing technology, communication technology, etc.,[2]. It has been a revolution of information sensing, collecting, and processing[3]. Micro-sensor technol- ogy supported by Micro-electro-mechanical systems and wireless communication ca- pabilities for sensor networks gives a broad application prospect for WSN[4]. WSN is composed of many low-power micro-sensors[5] whose nodes can sense various physi- cal phenomena in industrial, military, and other environments, such as sound, light, temperature, and movement. The nodes can process the original sensing data and trans- mit data to the sink node in wireless multi-hop routing [6]. The sink node sends the collected data to a wireless network or a local area network. It greatly improves the
2 accuracy and sensitivity of the data and information[7]. WSNs have the advantages of high precision monitor, strong fault tolerance, large coverage area, and capable of re- mote monitoring[8]. It has been widely used in military and national defense, national security, environmental monitoring, target tracking, medical and health, combating ter- rorism, intelligent building, environmental science and space exploration, and other fields [2][9]. Sensor nodes have limited capabilities due to small storage, low power due to battery equipped, communication cost, and limited processing capabilities[10]. Therefore, the sensor node is susceptible to node failure. This type of node failure causes the coverage hole in the network, and the sensed data cannot be transmitted between nodes [11]. Artificial bee colony (ABC) algorithm has high global optimality compared to other intelligent algorithms, and it is not easy to drop into the local optimum[12]. In the lo- calization process, however, there is an issue of slow convergence speed[13]. The ABC algorithm effectively enhances the honey source artificial bee's positioning efficiency and accuracy to update and calculate hired bees based on area limitation[14]. The initial honey source range significantly influences convergence speed and the artificial bee colony algorithm's optimization performance[15]. The random distribution and diver- sity of the population can be ensured to a certain degree if the random initialization approach is adopted. Still, the algorithm's convergence speed would be affected [16]. This paper focuses on the critical problem in WSNs, called the nodes coverage deploy- ment of WSN, and implementing an optimal sensor network based on the improved ABC algorithm. It means that the area coverage nodes and network lifetime is enhanced through optimizing the node deployment scheme in WSN by applying the improved ABC algorithm. 2 Related Works 2.1 Artificial bee colony algorithm The basic artificial bee colony algorithm (ABC) [12] is an intelligent algorithm to find the optimal solution through cooperation among individuals in the group. The global convergence of the algorithm is proved in reference [14]. The ABC is composed of three groups: employed bee, bystander bee, and reconnaissance bee[17]. The employed bees are used to mine the food sources, and the bystander bees randomly select a food source according to the probability of continuing to dig. The random food source loca- tion corresponds to the stochastic solution of the optimization problem, and the nectar quantity of the food source represents the fitness value. The number of employed wasps was the same as that of bystanders, while the number of scouts was only one. Suppose that in D-dimensional space, 𝑆𝑁 is the number of nectar sources and the location of i th food source 𝑋𝑖 = (𝑥𝑖1 , 𝑥𝑖2 , … , 𝑥𝑖𝐷 ), 𝑖 ≤ 𝑆𝑁. The process of ABC searching for the best food source includes the following steps. (1) Hire bee stage. Hired bees search the neighborhood around the current food source to generate new food sources, and then select a better food source according to greedy criteria. (2) The bee watching stage. According to the information shared by hired bees through
3 swing dance, bystander bees select food sources according to the probability of food source quality for neighborhood search. At the same time, a new food source is produced, and a better food source is selected according to the greedy selection mechanism. (3) The stage of bee detection. If a certain food source can not be further improved by the pre-set cycle number limit, it indicates that the food source has been exhausted. At this time, the employed bee will become the Scout bee, and the Scout bee will generate new food sources randomly. In the basic ABC algorithm, each worker and bystander will generate a new food source in the neighborhood of its current location through the following search equation: 𝑣𝑖𝑗 = 𝑥𝑖𝑗 + 𝜑𝑖𝑗 (𝑥𝑖𝑗 − 𝑥𝑘𝑗 ) (1) where j ∈ {1,2, … , 𝐷};i, k ∈ {1,2, … , 𝑆𝑁}, and i ≠ k , 𝜑𝑖𝑗 are random numbers in the range of [- 1,]. In the search process, after the hired bees share the food source infor- mation through the swing dance, the bystander bees select the food source according to the roulette selection mechanism. The probability value related to a food source is 𝐹𝑖𝑡𝑖 𝑃𝑖 = ∑𝑆𝑁 , (2) 𝑗=1 𝐹𝑖𝑡𝑗 where 𝐹𝑖𝑡𝑖 is the fitness value of food source i; the fitness calculation formula is as follows： 1 ,𝐹 ≥ 0 𝐹𝑖𝑡𝑗 = { (1 + 𝐹𝑖 ) 𝑖 (3) 1 + 𝑎𝑏𝑠(𝐹𝑖 ), 𝐹𝑖 < 0 where 𝐹𝑖 is the objective function value of corresponding food source i. If a food source is not updated within a limited period of time, the Scout bee will generate a new food source using the following formula: 𝑥𝑖𝑗 = 𝑙𝑗 + 𝑟𝑎𝑛𝑑(0,1)(𝑢𝑗 − 𝑙𝑗 ) (4) where 𝑙𝑗 and 𝑢𝑗 are the upper and lower bounds of dimension j, respectively. An itera- tive algorithm used to look for the best solution in the solution space is the ABC algo- rithm. When optimizing the node position problem, it can avoid the reliance on the initial value and has strong global optimality. However, the ABC algorithm also faces the problems of slow convergence speed and premature convergence, like other evolu- tionary algorithms, so it is hard to get optimal. 2.2 WSN Coverage Optimization Model Assuming that 𝑁 nodes are randomly deployed in the two-dimensional monitoring area, which is represented as 𝑁 = {𝑛1 , 𝑛2 , … , 𝑛𝑁 } and the coordinates of the node 𝑛𝑖 expressed by (𝑥𝑖 ,𝑦𝑖 ). The area is discretized into B×C grid points and the coordi- nates of each grid point is (x, y), 𝑥 ∈ 𝐵, 𝑦 ∈ 𝐶 [18]. The minimum number of nodes needed in the area 𝐴𝑎𝑟𝑒𝑎 is related to the sensing radius r of the nodes in [19]. The r is the maximum monitoring area that can be sensed by the node, and the value is usually determined by the hardware characteristics of the node itself. In addition, assuming that the communication radius of WSN nodes is larger than or equal to two times of the sensing radius r, in order to ensure the full connection of the network[8]. Therefore, the problem of coverage connectivity can be transformed into a different
4 coverage control problem. Based on the above assumptions, the perceptive duality model of grid point (x, y), covered by the WSN node, is: 1, 𝑖𝑓 𝑑 ≤ 𝑟 (5) 𝑝(𝑟) = { 0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒 where d is the distance from the WSN node to any grid point, which can be written as: (6) 𝑑 = √(𝑥 − 𝑥𝑖 )2 + (𝑦 − 𝑦𝑖 )2 When the grid point (x, y) is covered by a number of WSN nodes, the coverage prob- ability of the grid points is 1, regardless of the number of nodes. As a result, the total number of grid points covered by the target monitoring area is: (7) 𝑁𝑜𝑐𝑜𝑣𝑒 = ∑ 𝑝 𝐵×𝐶 Thus, the coverage rate of the WSN node in the target area Aarea can be written as: 𝑁𝑜𝑐𝑜𝑣 𝐴𝑎𝑟𝑒𝑎 = (8) 𝐵×𝐶 The target monitoring region Aarea have the same initial energy, i.e. the same initial sensing radius, and each node has the ability to self-regulate the sensing radius. The sensing radius set of nodes can be expressed as {𝑟1 , 𝑟2 , … , 𝑟𝑁 }, rmin ≤ ri ≤ rmax . The maximum sensing consumption of each node 2 is 𝑟𝑚𝑎𝑥 , and the maximum sensing 2 energy consumption 2 of the WSN network is 𝑁 − 𝑟𝑀𝑎𝑥 . The coverage optimization model is established based on maximizing the network coverage as the optimization function, while reducing the network sensing energy consumption is taken as the con- straint condition, by 𝑁𝑜𝑐𝑜𝑣 𝑀𝑎𝑥𝑖𝑚𝑖𝑧𝑒 𝐴𝑎𝑟𝑒𝑎 = 𝐵×𝐶 (9) 𝑠. 𝑡. ∑𝑖 𝑟𝑖2 < 𝑁 − 𝑟𝑀𝑎𝑥 2 The energy consumption of WSN is mainly embodied in three aspects: communica- tion energy consumption, computation energy consumption, and sensing energy con- sumption. In WSN coverage problem, it is focusing on the node's perception of energy consumption, less considering the communication, and calculating energy consump- tion. Assume that the energy consumption of the node in sleep state is 0. Let E and r denote the energy consumption and sensing radius of a node, respectively. The relation of E and r can be represented by Equation 𝐸 = 𝑘 × 𝑟2 (10) Usually, k is a constant greater than 0, typically k=1. By adjusting the sensing radius of WSN nodes, it can effectively improve network coverage and reduce sensing energy consumption.
5 3 Improved Artificial Bee Colony Algorithm for Nodes Coverage Deployment 3.1 Improving Artificial Bee Colony The region-restricted honey source updating is used to improve the ABC algorithm (namely IABC). Implementing the IABC algorithm aims to further enhance positioning efficiency, increase the convergence speed, and avoid the algorithm falling into local optimum in the late iteration process. To increase the diversity of honey sources in the late search period, avoid prematurely and falling into the ABC algorithm's local opti- mum, the disturbance frequency is increased in the process of position updating and the convergence speed. The location update formula is as follows: 𝑣𝑖𝑗 = 𝑥𝑖𝑗 + [𝑃𝑖𝑗 (𝑡) − 𝑟𝑎𝑛𝑑(0,1)] × (𝑥𝑖𝑗 − 𝑥𝑟𝑗 ) ∗ 𝛽 (11) where j ∈ {1,2, … , 𝐷};i, k ∈ {1,2, … , 𝑆𝑁}, D is a dimension, and i ≠ k, The position parameter is modified, a lower 𝛽 value would lead to a slow search and optimization process, while a higher 𝛽 value will increase the diversity of optimization solutions. The probability corresponding to the appearing position (x, y) bees can be updated by obtaining according to the Monte Carlo. 𝐿 𝑃𝑖𝑗 (𝑡) = 𝑃𝑎𝑣𝑔 (𝑡) ± 2𝐿𝑛(𝑢1 ) 𝑀 (12) 1 𝑃𝑎𝑣𝑔 (𝑡) = 𝑀 ∑ 𝑃𝑖𝑗 (𝑡) 𝑖=1,𝑗=1 (1−0.5)(𝑡𝑚𝑎𝑥 −𝑡) {𝛽 = 1⁄2 + 𝑡𝑚𝑎𝑥 where p(t) is the probability corresponding to the position (x, y); 𝑃𝑎𝑣𝑔 (𝑡) is the optimal location for bee agents; tmax is the maximum iterations number; u is a random number [0,1], if u≤0.5, the sign before β is "+", on the other hand, the symbol is "-". The pa- rameter β represents the velocity of bees convergence. 3.2 Energy Saving Coverage Optimization It is assumed that N nodes with sensing radius r are randomly deployed in the two- dimensional area of 𝐴𝑎𝑟𝑒𝑎 A, and the sensing region of the node is regarded as N disks. Discretization of the region into K grid points, and if the agent bee A is not covered, it will produce a gravitational force on its neighbor. I f the agent bee A has been covered, it will automatically shield disks. Thus, the quasi-gravitation expression of the disk i by the force of agent bee k, is [15] 𝑁 𝑟𝑖2 (13) 2 , 𝑣𝑘 ∉ ⋃ 𝑆𝑛 , 𝑅𝑖 ≥𝑑𝑖𝑘 >𝑟𝑖 𝑓𝑖𝑘 = {𝑑𝑖𝑘 𝑛=1 0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒 Among them 𝑑𝑖𝑘 indicates the distance between the bee's location and the center of the disk, 𝑟𝑖 is the weight of the disk; 𝑣𝑘 ∉ ⋃𝑁 𝑛=1 𝑆𝑛 is not the agent bees covered grav- itational effects on the disk, (he inequality 𝑟𝑖 < 𝑑𝑖𝑘 ≤ 𝑅𝑖 determines the scope of gravity. In addition to gravitation, a disk is also subjected to repulsion from other disks, which are called quasi coulomb forces. When the disk does not cover the agent bee
6 completely, the quasi coulomb force is in a failure state. When repulsion becomes the dominant force of moving disk, that is, all agent bee s are covered. This improves cov- erage performance by making the disc distribution more evenly and avoiding redundant repeat coverage. The quasi coulomb force is defined as follows. 𝑟𝑖2 2 , 𝑑𝑖𝑗 < min{𝑅𝑖 ,𝑅𝑗 } 𝑓𝑖𝑗 = { 𝑑𝑖𝑗 (14) 0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒 where dij is the distance between two nodes, the inequality 𝑑𝑖𝑗 < min{𝑅𝑖 , 𝑅𝑗 } limits the range of repulsive forces between neighbor nodes. The ABC algorithm is used to optimize the established coverage model based on maximizing the network coverage as the optimization function. 𝜔 × 𝑓𝑖𝑘 + (1 − 𝜔) × 𝑓𝑖𝑗 (15) 𝐹𝑢𝑛 = 𝐵×𝐶 Suppose the population size is 𝑆𝑁, in which the size of employed and observed bees is 𝑆𝑁 / 2. In the localization process, each honey source is defined as a two-dimen- sional coordinate vector, and the position of the nth honey source is recorded as 𝑋𝑛 = (𝑥𝑛1 , 𝑥𝑛2 ), 𝑛 ≤ 𝑆𝑁 where N is the number of anchor nodes. 𝐹𝑢𝑛 represents the objec- tive function value of the nth honey source under the u-th unknown node. The corre- sponding fitness values of nectar source were as follows: 1 𝐹𝑖𝑡𝑢𝑛 = { (1 + 𝐹𝑢𝑛 ) (16) 1 + 𝑎𝑏𝑠(𝐹𝑢𝑛 ) The formal description of the algorithm is as follows: S1. Initialize the population. At the same time, the fitness value of each honey source was calculated. The colony is made up of employment bees and observation bees. S2. Hired bees search for new honey sources near their respective honey sources. S3. The greedy selection strategy is used to determine whether to update the original honey source. S4. According to the greedy selection strategy, if the new honey source's fitness value is higher than that of the original honey source, the new position will be updated; oth- erwise, it will remain unchanged. S5. For each honey source, if the honey source's fitness value is not improved after several searches, the honey source will be abandoned and replaced by a new honey source randomly selected by the reconnaissance bee. S6. Record the optimal solution of each iteration, and repeat S2-S6 until the cycle con- ditions are met. 4 Simulation Results Suppose that the monitoring area is a square area of 𝐿 × 𝑊, e.g., 20m×20m (L=W=20, 30, 40, or 50 m) and disperse the region into a grid with a spacing of 2m, which forms 121 grid points. The network coverage and energy consumption are simulated by performing the proposed scheme 100 times and taking its average value. For convenience, we will present the coverage optimization algorithm as the improved
7 ABC algorithm. It is assumed that the initial sensing radius of each node r=2.5m, com- munication radius R=5m. The WSN nodes randomly generated in the monitoring area. a) The node's distribution of the ABC scheme b) The node's distribution of the ACO scheme c) The node's distribution of the PSO scheme d) The node's distribution of the IABC scheme Fig. 1. The outcomes of WSN node distributions of the proposed scheme based on the IABC compared with the PSO, ACO[20], and ABC[15] algorithms for the repetition coverage optimi- zation area. A consistent simulation condition is used in comparing the proposed suggestion scheme's efficiency. The simulation results of the proposed approach are compared with other methods for optimizing coverage nodes layouts with the mobile node loca- tion structure, such as the ABC[15] ant colony optimization algorithm (ACO) [20], and particle swarm optimization (PSO) [21]. The population size 𝑁𝑝 is set to 40 for the IABC, PSO, ACO, ABC algorithms; the number of iterations is set to 400; the dimen- sion is set to the number of mobile nodes; the random 𝑢 and probability 𝑝 are set to [- 1,1] and 0.1 respectively for IABC and ABC[15] algorithms. The weight 𝑤 is set to 04 to 0.9; 𝑐1 , 𝑐2 are set to 1.45; 𝑟1 , 𝑟2 are set to [0,1] for PSO[21]. The pheromone trails and visibility 𝛼 and 𝛽 are set greater than 0; the evaporation is set as [0,1] for the ACO[20]. Table 1. Four algorithms with different deployed network areas for the performance outcomes as the node layouts coverage optimization. Cov. No. PSO[21] ACO[20] ABC[15] IABC
8 area Mobile Cov. Iter. Cov. Iter. Cov. Iter. Cov. Iter. Nodes rate achieved rate achieved rate achieved rate achieved 15 × 15 12 79.7% 190 79.3% 199 78.0% 191 81.3% 209 25 × 25 15 81.0% 227 80.6% 240 79.4% 228 82.7% 254 35 × 35 20 83.5% 241 85.1% 255 81.8% 242 85.2% 271 40 × 40 25 85.9% 245 83.5% 260 82.2% 246 85.6% 275 45 × 45 30 85.1% 259 81.7% 275 80.4% 261 83.8% 232 50 × 50 35 81.0% 189 80.6% 198 82.4% 190 82.7% 208 Fig. 1 shows the outcomes of WSN node distributions of the proposed scheme based on the IABC compared with the PSO, ACO, and ABC algorithms for the repetition coverage optimization area. It can be seen that the node distribution in the coverage optimization area has some improvement after the implementation of the IABC algo- rithm. The node distribution is uneven, and there is still a lot of coverage of the blind area and repetition coverage area. Moreover, from Fig. 1, it is observed that the node distributions of the schemes, the IABC algorithm are more uniform, coverage blind area, and repetition coverage area are relatively small. Also, it can be seen that the node distribution of the IABC algorithm is better than that of the other algorithms. Table 2. Comparison of the proposed IABC approach's outcomes with the PSO, ACO, and ABC techniques for the coverage nodes error rate in WSN. IABC PSO[21] ACO[20] ABC[15] Methods Scheme Scheme Scheme Scheme Average coverage rate 83.43% 81.64% 80.52% 82.47% Executed time(second) 119.714 131.686 120.911 144.854 Table1 depicts the comparison of the proposed IABC approach's outcomes with the PSO, ACO, and ABC techniques for the coverage nodes in WSN. Table 2 compares the proposed IABC approach's outcomes in terms of the average node layout coverage and consumption time with the PSO, ACO, and ABC schemes for the coverage nodes in WSN. It can be seen that the suggested solution obviously has a higher degree of network coverage than the PSO, ACO, and ABC scheme approaches.
9 Fig. 2. Comparison the proposed IABC approach's outcomes in terms of the average node lay- out coverage with the PSO, ACO, and ABC schemes for the coverage nodes in WSN. The network coverage error rate comparison of the four schemes is given in Fig. 2. It can be seen that, when the proposed scheme based on the IABC algorithm converges to a stable coverage rate of 82.3%, the number of iterations is 250 times. In comparison, the ACB algorithm converges to a stable coverage rate of 75.9% with the iteration num- ber of 260 times. It shows that the global search ability of the IABC algorithm is better than the ABC algorithm. Also, it is observed that when the same coverage rate is achieved, the IABC scheme has a faster convergence speed. By adjusting the sensing radius dynamically, the IABC scheme can avoid the network redundancy to improve the coverage performance. Therefore, it concludes that the IABC algorithm has a faster convergence speed and better coverage performance. 5 Conclusion In this paper, we introduced an optimization node layout deployment scheme in Wire- less sensor networks (WSN), which is based on the improved artificial bee colony (IABC) algorithm to improve area coverage and network lifetime. The distance be- tween the sensor nodes is adjusted reasonably by implementing quasi-gravitation and quasi-coulomb power. Also, the sensing radius of WSN nodes is dynamically modified to enhance node energy consumption. A comparison between different algorithms com- monly used for network coverage shows that the proposed scheme increases area cov- erage and reduces the amount of energy consumed in WSN. The scheme of optimal layout coverage in WSN will be applied to the practical large scale WSN application in future work. References 1. Shiva Prasad Yadav, S.G., Chitra, A.: Wireless Sensor Networks - Architectures , Protocols , Simulators and Applications : a Survey. Int. J. Electron. Comput. Sci. Eng. 1, 1941–1953 (2012). 2. Nguyen, T.-T., Pan, J.-S., Chu, S.-C., Dao, T.-K., Do, V.-C.: Improved Performance of Wireless Sensor Network Based on Fuzzy Logic for Clustering Scheme. In: Advances in Smart Vehicular Technology, Transportation, Communication, and Applications. Springer, Cham, vol 128, 104-113 (2019). https://doi.org/10.1007/978-3-030-04585-2_13. 3. Nguyen, T.-T., Pan, J.-S., Dao, T.-K., Sung, T.-W., Ngo, T.-G.: Pigeon-Inspired Optimization for Node Location in Wireless Sensor Network BT - Advances in Engineering Research and Application. Presented at the (2020). 4. Dao, T., Nguyen, T., Pan, J., Qiao, Y., Lai, Q.: Identification Failure Data for Cluster Heads Aggregation in WSN Based on Improving Classification of SVM. IEEE Access. 8, 61070– 61084 (2020). https://doi.org/10.1109/ACCESS.2020.2983219. 5. Qiao, Y., Dao, T.K., Pan, J.S., Chu, S.C., Nguyen, T.T.: Diversity teams in soccer league competition algorithm for wireless sensor network deployment problem. Symmetry (Basel). 12, 445 (2020). https://doi.org/10.3390/sym12030445.
10 6. Nguyen, T.T., Pan, J.S., Dao, T.K.: An Improved Flower Pollination Algorithm for Optimizing Layouts of Nodes in Wireless Sensor Network. IEEE Access. 7, 75985–75998 (2019). https://doi.org/10.1109/ACCESS.2019.2921721. 7. Nguyen, T.T., Pan, J.S., Dao, T.K.: A compact bat algorithm for unequal clustering in wireless sensor networks. Appl. Sci. 9, (2019). https://doi.org/10.3390/app9101973. 8. Nguyen, T.-T., Dao, T.-K., Horng, M.-F., Shieh, C.-S.: An Energy-based Cluster Head Selection Algorithm to Support Long-lifetime in Wireless Sensor Networks. J. Netw. Intell. 01, 23–37 (2016). 9. Nguyen, T.-T., Pan, J.-S., Lin, J.C.-W., Dao, T.-K., Nguyen, T.-X.-H.: An Optimal Node Coverage in Wireless Sensor Network Based on Whale Optimization Algorithm. Data Sci. Pattern Recognit. 02, 11–21 (2018). 10. Thi-Kien Dao, Tien-Szu Pan, Trong-The Nguyen, Chu, S.-C.: A Compact Articial Bee Colony Optimization for Topology Control Scheme in Wireless Sensor Networks. J. Inf. Hiding Multimed. Signal Process. . 06, 297–310. 11. Chu, S.C., Dao, T.K., Pan, J.S., Nguyen, T.T.: Identifying correctness data scheme for aggregating data in cluster heads of wireless sensor network based on naive Bayes classification. Eurasip J. Wirel. Commun. Netw. 2020, 52(1–16) (2020). https://doi.org/10.1186/s13638-020-01671-y. 12. Karaboga, D., Basturk, B.: Artificial Bee Colony (ABC) Optimization Algorithm for Solving Constrained Optimization Problems. In: Foundations of Fuzzy Logic and Soft Computing. pp. 789–798 (2007). https://doi.org/10.1007/978-3-540-72950-1_77. 13. Karaboga, D., Gorkemli, B., Ozturk, C., Karaboga, N.: A comprehensive survey: Artificial bee colony (ABC) algorithm and applications. Artif. Intell. Rev. 42, 21–57 (2014). https://doi.org/10.1007/s10462-012-9328-0. 14. Dao, T.K., Pan, T.S., Nguyen, T.T., Chu, S.C.: A compact Articial bee colony optimization for topology control scheme in wireless sensor networks. J. Inf. Hiding Multimed. Signal Process. 6, 297–310 (2015). 15. Xu, X., Zhou, G., Chen, T.: Research on coverage optimisation of wireless sensor networks based on an artificial bee colony algorithm. Int. J. Wirel. Mob. Comput. 9, 199–204 (2015). 16. Nguyen, T.-T., Dao, T.-K., Kao, H.-Y., Horng, M.-F., Shieh, C.-S.: Hybrid Particle Swarm Optimization with Artificial Bee Colony optimization for topology control scheme in wireless sensor networks. J. Internet Technol. 18, (2017). https://doi.org/10.6138/JIT.2017.18.4.20150119. 17. Pan, T.-S., Dao, T.-K., Nguyen, T.-T., Chu, S.-C., Pan, J.-S.: Bees and pollens with communication strategy for optimization. (2016). https://doi.org/10.1007/978-3-662-49390- 8_63. 18. Nguyen, T.T., Shieh, C.S., Dao, T.K., Wu, J.S., Hu, W.C.: Prolonging of the network lifetime of WSN using fuzzy clustering topology. In: Proceedings - 2013 2nd International Conference on Robot, Vision and Signal Processing, RVSP 2013. pp. 13–16 (2013). https://doi.org/10.1109/RVSP.2013.10. 19. Dao, T., Yu, J., Nguyen, T., Ngo, T.: A Hybrid Improved MVO and FNN for Identifying Collected Data Failure in Cluster Heads in WSN. IEEE Access. 8, 124311–124322 (2020). https://doi.org/10.1109/ACCESS.2020.3005247. 20. Lee, J.-W., Lee, J.-J.: Ant-colony-based scheduling algorithm for energy-efficient coverage of WSN. IEEE Sens. J. 12, 3036–3046 (2012). 21. Li, W.: PSO based wireless sensor networks coverage optimization on DEMs. In: International conference on intelligent computing. pp. 371–378. Springer (2011).