## Xiaoling Guo , Xinghua Sun , Ling Li , Renjie Wu and Meng Liu## |

Parameter | Value |
---|---|

Region | 200 m×200 m |

Node number (N) | 200 |

Node initial energy | 0.5 J |

Base station coordinates | (100 m, 250 m) |

Cluster head selection probability (p) | 0.05 |

l packet size | 4000 bits |

Maximum iterations | 20 |

Population | 15 |

Maximum rounds | 1,500 |

To analyze the improved algorithm, we compare it with LEACH and CRISCA from three aspects: network life cycle, energy balance, and throughput.

5.2.1 Network life cycle

The green solid point is the cluster head node, the blue hollow circle is the member node, and the five-pointed star at the top is the base station. In Fig. 5, improved algorithm elects cluster heads according to the probability of 0.05. Before appearing dead node, the numbers of cluster head are fixed to be 10 in each round. Moreover, because base station is outside, the fitness function is constructed mainly based on inter-cluster distance. As can be seen from Fig. 5, cluster head is relatively close to the base station to avoid long distance transmission.

In Figs. 6 and 7, the dead nodes in LEACH and CRISCA are both at the bottom of region. These nodes die due to long distance transmission. Long distance transmission is not considered in the LEACH algorithm; in CRISCA, although the distance factor is taken into account when building the fitness function, the actual effect is not very ideal. In Fig. 8, the distribution of dead nodes is relatively uniform, because the improved algorithm constructs the fitness function with inter-cluster distance. In improved algorithm, cluster heads really play a role in avoiding long distance transmission.

Table 2.

Proportion of death nodes | LEACH | CRISCA | Improved algorithm |
---|---|---|---|

1st death node appears | 264 | 447 | 775 |

10% death nodes appear | 383 | 577 | 789 |

20% death nodes appear | 456 | 678 | 796 |

50% death nodes appear | 762 | 837 | 803 |

80% death nodes appear | 950 | 844 | 804 |

Occurrence of death nodes are greatly delayed in improved algorithm in Table 2 and Fig. 9. In LEACH the first dead node appears at 264th round, while it is at 775th round in improved algorithm, about 2.9 times than in LEACH. Moreover, the time of death of 10%, 20%, and 50% nodes in the improved algorithm are also delayed than that in LEACH and CRISCA. It only takes about 30 rounds from the first dead node to 80% dead nodes in improved algorithm. The 80% node death appears at 804th round in the improved algorithm, while that appears at the 950th and 844th round, respectively, in LEACH and CRISCA. Although there are some surviving nodes in them, there is a serious network segmentation in this region. Therefore, the significance of the existence of the network is not great.

In the paper, the distance between clusters is used to construct the fitness function, and the effect of distance on the balance of energy consumption is fully considered. In order to compare with the LEACH, CRISCA and the improved algorithm, the following scenarios are simulated in this paper, as shown in Table 3. Many simulation experiments have been carried out by changing the size of the monitoring area, the total number of nodes and the location of the base station. The average data of multiple runs are recorded. It can be seen that the first death node appearance in the improved algorithm occurs relatively late. The energy consumption is more balanced, and it only takes a short period from the first death node to 80% death nodes.

Table 3.

Region | Node number | Base station coordinates | Algorithm | Round of death nodes | |
---|---|---|---|---|---|

1st death node | 80% death nodes | ||||

200×200 | 200 | 100,250 | LEACH | 264 | 950 |

CRISCA | 447 | 844 | |||

Improved algorithm | 775 | 804 | |||

200×200 | 200 | 100,350 | LEACH | 103 | 519 |

CRISCA | 289 | 380 | |||

Improved algorithm | 337 | 391 | |||

100×100 | 100 | 50,150 | LEACH | 810 | 1157 |

CRISCA | 1002 | 1079 | |||

Improved algorithm | 1005 | 1021 | |||

100×100 | 100 | 50,250 | LEACH | 342 | 765 |

CRISCA | 389 | 659 | |||

Improved algorithm | 586 | 615 |

5.2.2 Energy balance

In Fig. 10, the death node of LEACH appears at about 260th round, that of CRISCA does at about 450th round, while that of improved algorithm appears much later, at about 770th round. Improved algorithm greatly prolongs the network life. Moreover, in improved algorithm, as soon as the first dead node appears, 80% of the nodes die quickly. It well balances energy consumption among nodes. About 80% of the dead nodes of LEACH algorithm appears at about 950th round. And 80% of the death nodes of CRISCA appears about at 844th round. The reason is that nodes far from base station die prematurely, thus remaining surviving nodes become closer to base station, resulting in delaying death time of these surviving nodes.

5.2.3 Throughput

In Fig. 11, the packets sent by improved algorithm is more than those by LEACH and CRISCA. Because appearing time of dead node in improved algorithm is delayed, survival time of whole network is prolonged, the time that all nodes can gather and transmit data becomes longer, and the total amount of data packets sent gets more. In addition, LEACH and CRISCA can continue sent data to base station after about 800th round, since there are some surviving nodes gathering and transmitting data in them.

From the perspective of prolonging network life cycle and balancing energy consumption, a new centralized clustering algorithm is proposed based on SCA in this paper. The algorithm determines the optimal number of cluster heads according to the current surviving node to adapt to the dynamic changes in network size. And the algorithm uses high-energy nodes to construct candidate cluster heads, and constructs fitness function according to the distance between clusters. Then, the algorithm uses SCA based on monotone decreasing convex function to find the optimal cluster head scheme with the lowest communication cost. The algorithm can well balance energy consumption among nodes to improve network energy utilization efficiency and throughput. And it can greatly postpone the rounds of death nodes to extend the effective life of the whole network. Moreover, compared with CRISCA algorithm, the fitness function of algorithm is much simpler and more efficient, and the burden on the base station gets lighter. The algorithm has better prospects in large-scale networks, but there are still some areas to be improved. In the following step, we will further study the layering or chain algorithm between clusters in a larger scale network deployment scenario.

He is working as a professor in School of Information Science and Engineering in Hebei North University, Zhangjiakou, China. He received M.S. degree in School of Mathematical Information in Shanghai Normal University in 2007. His current research interests include mobile communication and lighting control network.

She is currently working at the College of Information Science and Engineering, Hebei North University, Zhangjiakou, China. She received her master's degree in management from Beijing University of Chinese Medicine in 2020. Her main research interests are computer networks and information management.

- 1 W. R. Heinzelman, A. Chandrakasan, and H. Balakrishnan, "Energy-efficient communication protocol for wireless microsensor networks," in
*Proceedings of the 33rd Annual Hawaii International Conference on System Sciences*, Maui, HI, 2000, pp. 3005-3014.doi:[[[10.1109/hicss.2000.926982]]] - 2 D. Li and D. M. Xu, "Improvement of LEACH algorithm in wireless sensor networks,"
*Computer Engineering and Design*, vol. 41, no. 7, pp. 1852-1857, 2020.doi:[[[10.1109/SENSORCOMM.2007.4394931]]] - 3 Y . F. Tian and L. H. Wang, "Routing algorithm for wireless sensor networks by considering residual energy and communication cost,"
*Journal of Nanjing University of Science and Technology*, vol. 42, no. 1, pp. 96101, 2018.custom:[[[-]]] - 4 X. H. Li, Y . C. Zhao, and J. P . Zhao, "Poisson distribution optimal number of cluster heads in wireless sensor networks,"
*Communications Technology*, vol. 53, no. 2, pp. 335-340, 2020.custom:[[[-]]] - 5 H. B. Li, Z. L. Liu, Q. Chen, S. Liu, X. L. Liu, Y . Q. Liang, Z. Yang, and L. W. Chen, "Clustering routing algorithm for wireless sensor network based on hierarchical neighboring nodes,"
*Computer Engineering*, vol. 46, no. 6, pp. 187-195, 2020.doi:[[[10.19678/j.issn.1000-3428.0054723]]] - 6 J. Kim, D. Lee, J. Hwang, S. Hong, D. Shin, and D. Shin, "Wireless sensor network (WSN) configuration method to increase node energy efficiency through clustering and location information,"
*Symmetry*, vol. 13, no. 3, article no. 390, 2021. https://doi.org/10.3390/sym13030390doi:[[[10.3390/sym13030390]]] - 7 Z. S. Wang, H. W. Ding, B. Li, H. Li, and A. S. Li, "Energy-efficient WSNs routing protocol based on clustering,"
*Computer Engineering and Design*, vol. 42, no. 2, pp. 324-330, 2021.custom:[[[-]]] - 8 P . Maheshwari, A. K. Sharma, and K. V erma, "Energy efficient cluster based routing protocol for WSN using butterfly optimization algorithm and ant colony optimization," Ad Hoc Networks, vol. 110, article no. 102317, 2021. https://doi.org/10.1016/j.adhoc.2020.102317doi:[[[10.1016/j.adhoc..102317]]]
- 9 S. Sefati, M. Abdi, and A. Ghaffari, "Cluster-based data transmission scheme in wireless sensor networks using black hole and ant colony algorithms,"
*International Journal of Communication Systems*, vol. 34, no. 9, article no. e4768, 2021. https://doi.org/10.1002/dac.4768doi:[[[10.1002/dac.4768]]] - 10 N. Ajmi, A. Helali, P . Lorenz, and R. Mghaieth, "MWCSGA—Multi weight chicken swarm based genetic algorithm for energy efficient clustered wireless sensor network," Sensors, vol. 21, no. 3, article no. 791, 2021. https://doi.org/10.3390/s21030791doi:[[[10.3390/s0791]]]
- 11 X. He, Z. Z. Ning, and X. Y . Yang, "WSN clustering routing protocol based on improved sine cosine algorithm,"
*Journal of Xi'an University of Post and Telecom*, vol. 26, no. 2 pp. 15-20, 2021.custom:[[[-]]] - 12 S. Mirjalili, "SCA: a sine cosine algorithm for solving optimization problems,"
*Knowledge-based Systems*, vol. 96, pp. 120-133, 2016.doi:[[[10.1016/j.knosys.2015.12.022]]] - 13 C. Chen, L. Ma, and Y . Liu, "Quantum sine cosine algorithm for function optimization,"
*Application Research of Computers*, vol. 34, no. 11, pp. 3214-3218, 2017.custom:[[[-]]] - 14 L. Q. Yong, Y . H. Li, and W. Jia, "Literature Survey on Research and Application of Sine Cosine Algorithm,"
*Computer Engineering and Applications*, vol. 56, no. 14, pp. 26-34, 2020.custom:[[[-]]] - 15 Y . Liu and L. Ma, "Sine cosine algorithm with nonlinear decreasing conversion parameter,"
*Computer Engineering and Applications*, vol. 53, no. 2, pp. 1-5, 2017.doi:[[[10.3778/j.issn.1002-8331.1608-0349]]] - 16 D. Chen and Y . H. Zhao, "Speed control of permanent magnet synchronous motors using fuzzy PI controller based on sine cosine algorithm,"
*Electric Drive*, vol. 49, no. 5, pp. 31-36, 2019.doi:[[[10.1016/j.ijepes.2012.06.013]]] - 17 P . S. Rao, P . Lalwani, H. Banka, and G. S. N. Rao, "Competitive swarm optimization based unequal clustering and routing algorithms (CSO-UCRA) for wireless sensor networks,"
*Multimedia Tools and Applications*, vol. 80, no. 17, pp. 26093-26119, 2021.doi:[[[10.1007/s11042-021-10901-4]]] - 18 S. P . Su and H. j. Zhang, "Scheme of adaptive clustering based on optimal number of cluster nodes for WSNs,"
*Control Engineering of China*, vol. 24, no. 5, pp. 1070-1074, 2017.custom:[[[-]]] - 19 K. S. Arikumar, V . Natarajan, and S. C. Satapathy, "EELTM: an energy efficient LifeTime maximization approach for WSN by PSO and fuzzy-based unequal clustering,"
*Arabian Journal for Science and Engineering*, vol. 45, no. 12, pp. 10245-10260, 2020.doi:[[[10.1007/s13369-020-04616-1]]]