2019 Workshop on Computing, Networking and Communications (CNC)

On Cluster Head Selection in Monkey-inspired

Optimization based Routing Protocol for WSNs
Tina Gui, Feng Wang, Christopher Ma and Dawn E. Wilkins
Department of Computer and Information Science
University of Mississippi
University, MS 38677, USA
Email: {tgui, cma}@go.olemiss.edu; {fwang, dwilkins}@cs.olemiss.edu

Abstract—Wireless Sensor Networks (WSNs) have arisen as
one of the hottest and most promising research areas nowadays,
where one most critical research challenge is to extend the
network lifetime, as in most cases sensor nodes can hardly be
replaced with new batteries or recharged regularly. As a result,
WSNs are often divided into clusters to reduce communication
cost and energy consumption for packet routing, and the selection
of cluster head can thus have great impacts on the overall
network lifetime citeyang2018researchṘecently, monkey-inspired
optimization (MO), as one of the latest members to the Swarm
Intelligence (SI) family, has claimed its dedication to design
smart, cost efficient yet simple and deployable clustering routing
protocols for WSNs. In this paper, we conduct a detailed
investigation on cluster head selection for such MO based routing
protocols. We then propose an effective cluster head selection
scheme based on the state-of-the-art Spider Monkey Optimization
formulation, named SMOCH, and apply a greedy selection
process between existing node positions and newly generated
positions. Our experiment results show that our cluster-head
selection scheme can observably increase node lifetime to a larger
number of rounds and also decrease the energy consumption for
communications.
Index Terms—Wireless Sensor Networks, Swarm Intelligence,
Monkey-Inspired Optimization, Routing Protocol
I. I NTRODUCTION
Wireless Sensor Networks (WSNs) have arisen as one of the
hottest and most promising research areas nowadays, where
one of the most critical research issues is to extend the
network lifetime, as in most cases sensor nodes can hardly be
replaced with new batteries or recharged regularly. As a result,
WSNs are often divided into clusters to reduce communication
cost and energy consumption for packet routing, and the
selection of cluster head can thus have great impacts on the
overall network lifetime [1]-[2]. Recently, monkey-inspired
optimization (MO) has been introduced as one of the latest
members to the Swarm Intelligence (SI) family. At present,
scientists have discovered over 260 species that fit the
definition of monkey. A new finding designates that many
aspects of the human brain that control thought processes
are similar to those of monkeys. In ethology, a fission-fusion
society represents a group of animals change over time, where
they split in small groups (fission) during the day and sleep
together (fusion) at night. Spider monkeys are a good example
of this fission-fusion society. Typically, the female monkey
has higher rank than the males and acts like a leader. Male
monkeys are less dominant but more energetic. In this context,
the swarm intelligence in monkeys have been well studied,
yielding a series of most newly developed algorithms in
the family of optimizations inspired by animals and insects,
where some of the famous family members include Particle
Swarm Optimization (PSO), Ant Colony Optimization (ACO),
Artificial Bee Colony (ABC), Termite Colony Optimization
(TCO), and Spider Monkey Optimization (SMO). As shown
in Figure 1, the attracted research efforts in Swarm Intelligence
field have exploded exponentially.
Fig. 1. Number of publications related to swarm intelligence [3].
In the literature, Spider Monkey Optimization (SMO) is a
new meta-heuristic algorithm that falls in the discipline of
Swarm Intelligence. This is a collection of social behaviors
of spider monkeys that coordinate using self-organization
and decentralized control. This technique has demonstrated
the desirable properties of robustness, interpretability,
effectiveness and scalability [4]. Spider monkey optimization
has been widely used in various applications such as
electric power systems [5], numerical optimization [6], image
segmentation [7], and big data [8]. It recently has also claimed
its dedication to design smart, cost efficient yet simple and
deployable clustering routing protocols for WSNs [9].
In this paper, we conduct a detailed investigation on cluster
head selection for such MO based routing protocols. In
general, a good cluster head candidate should be a sensor
node that has the maximum number of neighboring nodes, the
shortest distance to the base station and the highest residual
energy. Yet in practice, it is usually hard to simultaneously
achieve all these requirements due to the involved optimization
complexity. To this end, we propose an effective cluster

978-1-5386-9223-3/19/$31.00 ©2019 IEEE 126

 2019 Workshop on Computing, Networking and Communications (CNC)
head selection scheme (SMOCH) based on the state-of-the-
art Spider Monkey Optimization formulation, and apply a
greedy selection process between existing node position and
newly generated position. In SMOCH, local data fusions and
minimum energy transmissions are performed to send data
from the clusters to the base station with the optimal routes,
which further ensures more nodes can stay alive and reduces
energy dissipation.
As comparing different routing protocols in WSNs can be
a challenging task. We also strive to propose generalized
performance metrics for evaluating MO based routing
protocols with different cluster head selection schemes, realize
potential synergies, , where our evaluation show that SMOCH
scheme can observably increase node lifetime to a larger
number of rounds and also decrease the energy consumption
for communications. Based on the results, we also identify
possible issues for future investigation and improvement.
The rest of the paper is organized as follows. The existing
weakness and strengths of the SI based routing protocols
are addressed in Section II. In Section III, we provide a
brief discussion on the related work in terms of cluster head
selection techniques for WSNs and present our SMOCH
scheme therein. Section IV describes the simulation model
and presents binarytree the experimental results of SMOCH
and the comparison with other MO based protocols. Finally,
we conclude the paper and discuss about future directions in
Section V.
II. D ESIGN C HALLENGES
Saleem et al. [10] proposed a general list of essential factors
for wireless sensor network routing protocols like scalability,
self-organization, memory requirements, sensor localization,
fault tolerance, energy efficiency and security. Due to the
large amount of sensor nodes, unstable energy sources
and unpredict operating environment, these facts present
unique challenges on the architecutral design and application
development of WSNs. Particularly, major concerns in
routing protocols developed by monkey-inspired optimization
are robustness, reachability, scalability, simplicity, coverage,
routing strategies, flexibility and quality of service. In this
section, a detailed explanation on some of these factors is
provided as below:
1) Scalability: it is a challenging problem in WSNs –
large amount of sensor nodes are expected to be
heterogeneously deployed with long transmission path
[10].
2) Self Organization: monkey inspired protocols must be
flexible to unpredict variations, in static and dynamic
manners [11].
3) Multipath Routing: adopting multiple path strategy can
extend network lifetime. It can be considered as a
backup when the initial path fail [12].
4) Localized Interaction: full locality awareness, restrition
of interactions to neighboring sensor nodes, is
a main charactertic of self-organization. Types of
127
communication in WSNs are aggregation, distribution,
and broadcast [13].
5) Failure Detection: the communication among sensor
nodes can be affected by battery life, location, weather
condition, obstacles, antenna and etc [14].
6) Memory Requirements: Given limited on board memory,
the routing algorithms developed require minimal
processing overhead in order to have an efficient
execution of functionalities [15].
7) Energy Efficiency: Due to limited battery life on
sensor nodes, the power usage effectiveness is a
critical challenge in WSNs to support longer network
operational time [15].
8) Robustness: Each sensor node needs to be constructed as
robust as possible in case of battery life shortening. The
robustness can be improved through the use of multi-
channel and multi-radio [16].
III. MO- BASED ROUTING P ROTOCOLS
A. Rhesus Macaque Optimization (RMO) based Protocol
Sandeep et al. [17] introduced an embeded bio-inspired
scheme into the LEACH routing protocol [18] for clustered
sensor networks. Rhesus macaque is one of the well-known
of Old World Monkeys. In contrast to the traditional routing
protocols, Rhesus Macaque Optimization algorithm is adopted
at the set up phase by reducing the overall energy consumed
in the cluster head selection step. The authors proposed three
key attributes taken into account for the design of the protocol
to improve network lifetime which are summarized as follows:
1) Election of cluster head (CH) is based on the residual
energy of the sensor nodes. Similar social behavior is
observed in monkey cubs, where the female monkey is
always selected as group leader according to her fitness
and fertility.
2) In monkey groups, the leader is surrounded by 80% of
the members. If the group size increases, it will split into
smaller groups and choose new subgroup leaders. This
behavior is also applied to reducing energy consumption
in sensor networks. The newly elected cluster heads
facilitate good communication within its member nodes
and transfer data packets to other active cluster heads.
3) If the group leader passes away, the second highest
ranked female monkey will replace her; and this process
of inheritance will last until there are not enough
members in the group. Similarly, the cluster head
without any sensor nodes attached to it is considered a
redundant cluster head. It loses its designation and will
be adopted by the nearest cluster.
The process of eliminating cluster heads and re-attaching to
a new group mimics the social behavior of a male monkey, as
shown in Fig. 2. The process of breaking a sensor cluster is
illustrated in Fig. 3. The RMO algorithm adopts similar cluster
head selection strategy as used in the LEACH protocol. The
cluster head can be selected stochastically, which is shown as
follows:
 2019 Workshop on Computing, Networking and Communications (CNC)

Algorithm 1 cluster head selection for RMO based routing dissipation in WSNs. Fig. 4 shows an illustration of how
protocol [17] SMOCH works, where it updates the position of individuals
Set-up Phase based on their fitness scores to increase the convergence
1: Construction of sensor network performance and exploitation capability. In this way a better-
2: Run the LEACH model once positioned node can have a higher chance to be chosen as
3: Find redundant cluster heads cluster head. SMOCH extends the ability of the original Spider
4: Attach redundant cluster heads to the nearest cluster Monkey Optimization algorithm and LEACH by nominating
Steady State the cluster heads with better location in order to improve
5: Randomly choose nodes for communication overall network performance in terms of less dead nodes and
energy consumption. The procedure of position update process
in SMOCH algorithm is shown as below:

Algorithm 2 cluster head selection for SMO based routing

protocol
Initialize population with n spider monkeys SM, k random
cluster centers and cluster head probability CHprob
1: while max number of iteration is not reached do
2: for each SM do
3: Calculate Euclidean distance of SM with all cluster
Fig. 2. RMO: Cluster Head Redundant and Re-attach Process [17] centroids
4: Assign SM to the cluster that has the nearest centrism
5: end for
6: Calculate the fitness and CHprob
7: if U (0, 1) < CHprob then
8: Update the cluster centroids based on local best
9: end if
10: end while

Fig. 3. RMO: The Process of Breaking the Clusters [17]

B. Spider Monkey Optimization (SMO) base Protocol and

SMOCH
Spider monkey optimization (SMO) is a relative new
addition to the family of swarm intelligence algorithms by
Fig. 4. SMO: The Process of New Cluster Heads Election
structuring the social foraging behavior of spider monkeys. In
our previous research [9], we aimed to study the mechanism
of SMO in the routing optimization of WSNs, formulating IV. E XPERIMENTAL R ESULTS AND D ISCUSSIONS
a mathematical model of the behavior patterns on which the
Spider Monkey Optimization (SMO-C) approach is adapted A. Evaluation Environment
and sending information from the clusters to the base station
through the optimal route. Additionally, our methodology In our evaluation environment, we simulated each monkey
based on the Spider Monkey’s behavioral structure has inspired optimization technique with MATLAB by randomly
improved the traditional routing protocols in terms of less distributing all sensor nodes in an area of 100m * 100m.
energy consumption and better service quality of the network. The base station, which acts like a gateway between sensor
In this paper, we present an effective cluster head selection nodes and the end user, is placed in a permanent location in
scheme (SMOCH) combining with our previous SMO-C the sensing area. Experimental parameters are summarized in
protocol to maximize network lifetime and minimize energy TABLE I

128
 2019 Workshop on Computing, Networking and Communications (CNC)

TABLE I
E XPERIMENTAL S ETTINGS

Parameter Symbol Value

Number of nodes n 100
Number of iterations r 500, 1000
Base station (x, y) (150, 50)
p
Distance threshold d (Ef s /Eamp )m
Percentage of cluster heads P 20%
Transmitter electronics Etx 50 nJ/bit Fig. 5. Average Remained Energy of RMO and SMOCH (Einit = 0.1J/node
Receiver electronics Erx 50 nJ/bit for 100 nodes after 500 and 1000 rounds).
Transmit Amplifier type 1 Ef s 10 pJ/bit/m2
Transmit Amplifier type 2 Eamp 0.0013 pJ/bit/m2
Data aggregation energy Eda 5 nJ/bit
Initial energy Einit (0.1, 0.5, 1.0) J/node

B. Results and Discussions

In LEACH approach, the cluster heads are unevenly

distributed and RMO is mainly designed based on LEACH Fig. 6. Average Remained Energy of RMO and SMOCH (Einit = 0.5J/node
for 100 nodes after 500 and 1000 rounds).
strategy. In some sensing areas, the nodes are positioned
at a great distance from the cluster heads and cannot be
replaced or recharged regularly. To improve the full network
performance such situations have been considered in SMOCH,
where the sensor nodes elevate themselves as cluster heads
according to the state-of-the-art Spider Monkey Optimization
formulation. The associated cluster heads are located within
the structured local clusters with best fitness scores. SMOCH
is a decentralized, self-organizing and interpretable clustering
protocol. Moreover, SMOCH performs local data fusion and
minimum energy transmission, further enhancing more alive Fig. 7. Average Remained Energy of RMO and SMOCH (Einit = 1.0J/node
nodes and reducing energy dissipation [9]. for 100 nodes after 500 and 1000 rounds).
Figures 5-7 indicate the average remained energy of each
node in SMOCH is greater than RMO in each round. It
1) Easy to design and interpret: it is a non-parameteric
also can be inferred from Figures 8-10 that the duration of
optimization algorithm which means no manual
most sensor nodes is prolonged in the proposed SMOCH
parameter setting required [20].
and it has the lowest number of dead nodes in the long
2) Capability in multi-path model: swarms are divided into
run. The sensor nodes can last longer when given limited
multiple groups; all groups exchange information and
initial energy and extend the overall network lifetime. During
intelligence to optimize the routing solutions [21].
the cluster head selection step, we achieve better results
3) Self-organization behavior: all swarms act at the same
compared with Rhesus Macaque Optimization (RMO). Our
time and there is no central coordinator [4].
simulation outcomes show that with improved cluster-head
locations, we can observably decrease the energy consumption
for communication and increase the network lifetime in a V. C ONCLUSION AND F UTURE D IRECTIONS
larger number of rounds. Contrasting with the other clustering Wireless sensor networks have broad impacts in many
protocols, SMOCH and RMO protocols focus on decreasing applications, such as battlefield surveillance, security,
the energy consumed at the set-up phase not the steady phase. healthcare, natural disaster and environmental monitoring.
The experimental results show that the proposed SMOCH The development of cost efficient, reliable and easily
approach is self-organized, scalable and can be easily adapted deployable sensor networks is challenging. In this work, we
in the wireless sensor networks. discussed generalized performance metrics for comparing
The monkey-inspired optimization algorithms require less the MO based routing protocols with different cluster head
computation time than conventional algorithms [19]. Our selection schemes, realized potential synergies, and identified
proposed SMOCH approach improves the exploration and possible issues for future investigation and improvement.
exploitation capabilities of the search space. We believe the Our experimental results show that with better cluster-head
SMO based routing protocol will provide better performance locations, our proposed SMOCH can observably increase
due to the following reasons: the nodes’ lifetime in a larger number of rounds and

129
 2019 Workshop on Computing, Networking and Communications (CNC)
Fig. 8. Number of Dead Nodes of RMO and SMOCH (Einit = 0.1J/node
for 100 nodes after 500 and 1000 rounds).
Fig. 9. Number of Dead Nodes of RMO and SMOCH (Einit = 0.5J/node
for 100 nodes after 500 and 1000 rounds).
Fig. 10. Number of Dead Nodes of RMO and SMOCH (Einit = 1.0J/node
for 100 nodes after 500 and 1000 rounds).
decrease the energy loss of communication. In addition, as
in monkeys’ social behaviors, they do not maintain fixed
size of subgroups through the foraging process, means the
percentage of choosing cluster heads in routing protocols
will not affect the overall performance. This observation
needs to be examined in our future work. We hope that
our work will trigger broader discussions and attract more
sensor network engineers and researchers to take into account
the various monkey-inspired optimization algorithms and
implementations in next generation of WSN applications.
R EFERENCES
[1] A.-M. Yang, X.-L. Yang, J.-C. Chang, B. Bai, F.-B. Kong, and Q.-B.
Ran, “Research on a fusion scheme of cellular network and wireless
sensor for cyber physical social systems,” IEEE Access, vol. 6, pp.
18 786–18 794, 2018.
[2] Y. Liu, Y.-J. Zhan, and J. Chen, “An integrated approach to sink
and sensor role selection in wireless sensor networks: Using dynamic
programming.” Adhoc & Sensor Wireless Networks, vol. 22, 2014.
[3] “Web of science,” https://clarivate.com/products/web-of-science,
Clarivate Analytics, accessed: 2018-05-30.
[4] T. Gui, C. Ma, F. Wang, and D. E. Wilkins, “Survey on swarm
intelligence based routing protocols for wireless sensor networks:
An extensive study,” in Industrial Technology (ICIT), 2016 IEEE
International Conference on. IEEE, 2016, pp. 1944–1949.
130
[5] N. Nayak, M. S. Mahali, I. Majumder, and R. K. Jena, “Dynamic
stability improvement of vsc-hvdc connected multi machine power
system by spider monkey optimization based pi controller,” in Electrical,
Electronics, and Optimization Techniques (ICEEOT), International
Conference on. IEEE, 2016, pp. 152–157.
[6] J. C. Bansal, H. Sharma, S. S. Jadon, and M. Clerc, “Spider monkey
optimization algorithm for numerical optimization,” Memetic computing,
vol. 6, no. 1, pp. 31–47, 2014.
[7] S. S. Pal, S. Kumar, M. Kashyap, Y. Choudhary, and M. Bhattacharya,
“Multi-level thresholding segmentation approach based on spider
monkey optimization algorithm,” in Proceedings of the Second
International Conference on Computer and Communication
Technologies. Springer, 2016, pp. 273–287.
[8] S. Sriprasadh, “Spider monkey optimization for data manipulation in big
data.” nternational Journal of Control Theory and Applications, vol. 9,
2016.
[9] T. Gui, C. Ma, F. Wang, J. Li, and D. E. Wilkins, “A novel cluster-
based routing protocol wireless sensor networks using spider monkey
optimization,” in Industrial Electronics Society, IECON 2016-42nd
Annual Conference of the IEEE. IEEE, 2016, pp. 5657–5662.
[10] M. Saleem, G. A. Di Caro, and M. Farooq, “Swarm intelligence
based routing protocol for wireless sensor networks: Survey and future
directions,” Information Sciences, vol. 181, no. 20, pp. 4597–4624, 2011.
[11] A.-S. K. Pathan, H.-W. Lee, and C. S. Hong, “Security in wireless
sensor networks: issues and challenges,” in Advanced Communication
Technology, 2006. ICACT 2006. The 8th International Conference,
vol. 2. IEEE, 2006, pp. 6–pp.
[12] I. Kassabalidis, M. El-Sharkawi, R. Marks, P. Arabshahi, and A. Gray,
“Swarm intelligence for routing in communication networks,” in Global
Telecommunications Conference, 2001. GLOBECOM’01. IEEE, vol. 6.
IEEE, 2001, pp. 3613–3617.
[13] Y. Liu, J. Chen, and Y.-j. Zhan, “Local patches alignment embedding
based localization for wireless sensor networks,” Wireless personal
communications, vol. 70, no. 1, pp. 373–389, 2013.
[14] H. Liu, A. Nayak, and I. Stojmenović, “Fault-tolerant
algorithms/protocols in wireless sensor networks,” in Guide to
Wireless Sensor Networks. Springer, 2009, pp. 261–291.
[15] A. Datta and S. Nandakumar, “A survey on bio inspired meta
heuristic based clustering protocols for wireless sensor networks,” IOP
Conference Series: Materials Science and Engineering, vol. 263, no. 5,
2017.
[16] P. Cui, Y. Dong, H. Liu, D. Rajan, E. Olinick, and J. Camp, “Whitemesh:
Leveraging white spaces in wireless mesh networks,” in Modeling and
Optimization in Mobile, Ad Hoc, and Wireless Networks (WiOpt), 2016
14th International Symposium on. IEEE, 2016, pp. 1–7.
[17] S. Kumar and S. Kusuma, “Clustering protocol for wireless sensor
networks based on rhesus macaque (macaca mulatta) animal’s social
behavior,” International Journal of Computer Applications, vol. 87,
no. 8, 2014.
[18] W. R. Heinzelman, A. Chandrakasan, and H. Balakrishnan, “Energy-
efficient communication protocol for wireless microsensor networks,”
in System sciences, 2000. Proceedings of the 33rd annual Hawaii
international conference on. IEEE, 2000, pp. 10–pp.
[19] R. Devi and S. Sathya, “Monkey behavior based algorithms-a survey,”
International Journal of Intelligent Systems and Applications, vol. 9,
no. 12, 2017.
[20] S. Kumar, V. K. Sharma, and R. Kumari, “Self-adaptive spider monkey
optimization algorithm for engineering optimization problems,” JIMS8I-
International Journal of Information Communication and Computing
Technology, vol. 2, no. 2, pp. 96–107, 2014.
[21] J. Dhar and S. Arora, “Designing fuzzy rule base using spider monkey
optimization algorithm in cooperative framework,” Future Computing
and Informatics Journal, vol. 2, no. 1, pp. 31–38, 2017.