Clustering is the most popular strategy for increasing the lifetime of a wireless sensor network, which immediately leads to a stronger routing process. This process requires the processing of sensor nodes into clusters and assigning relevant cluster heads to each cluster. This paper aims to implement a new hybrid algorithm called Over taker Assisted Political Update (OA-PU) for selecting an efficient cluster head. This cluster head is selected based upon four factors, namely energy, distance, cluster radius, and time. The hybrid algorithm is a combination of the Fusion Rider Optimization Algorithm (F-ROA) as well as a nature-inspired Political Optimization Algorithm (PO). Fusion-ROA depends on groups of riders that are attempting to reach the goal and PO is a state of the art meta-heuristic optimization methodology for global and structural analysis performance prediction in the WSN environment. In terms of alive node analysis, cost function analysis, and energy analysis, the proposed OA-PU outperforms conventional approaches. Consequently, the lifetime of WSN is increased and an optimal system is developed.
The network lifespan is considered a valid indicator of network efficiency in wireless sensor networks (WSN). The network runtime is defined by the time it takes for a particular amount of nodes to perish source of power exhaustion [
Behavioral and physiological factors may be sensed and tracked using numerous sensor nodes (SNs) in the WSN. The evolution of sensors has made WSN currently available in a compact SN format with wireless networking hardware [
The local search technique obtained through Cuckoo Search Optimization Algorithm (CSOA), is also insufficient. Thus, for increasing network lifetime and managing resources in the network, the discovery constraint of GWOA and the exploitation limitations of CSOA under cluster head selection must be efficiently managed [
Overall, the interval between member nodes as well as their respective cluster heads is short, and intra-cluster contact is accomplished with a single hop. The gap between the cluster heads and the base station is usually very large. If single-hop connectivity is used, the cluster head’s energy utilization would be increased unnecessarily. The distance between the cluster head as well as the base station in large-scale WSNs can also surpass the sensor node’s full radio range [
Sensor nodes in the same cluster can interact directly with each other, dividing the wireless sensor network into clusters. Based on the intensity of received signals, a sensor node selects which cluster to enter [
Cluster head selection is a difficult process as it has several problems including energy consumption, reduced lifespan, data transmission, execution time, and residual energy. By reducing the cluster head selection time the energy consumption can be reduced and thereby increasing the life span of WSN. Henceforth, a hybrid algorithm called Over Taker Assisted Political Update (OA-PU) is developed by combining the concepts of Fusion Rider Optimization Algorithm (F-ROA) and nature-inspired Political Optimization Algorithm (PO) to reduce energy utilization and increase network life-cycle. To choose a CH, parameters like energy, distance, cluster radius, and time are used. Hereby, the OA-PU ensures that CHs delivered aggregate data successfully to the Base Station (BS). The suggested model maintains a high coverage ratio and eliminating redundancy. The threshold value of residual energy is calculated using our current method. In terms of alive socket analysis, decision variables analysis, and conceptual design, the functionality of the OA-PU is contrasted to that of standard techniques.
The literature works related to our research are reviewed in the second section. Section three describes the network and energy consumption models of our proposed Optimization scheme. The proposed algorithm is simulated and compared to other existing methods in the fourth section. Section six concludes the paper with the obtained results and outlines the future scope of this research work.
Pitchaimanickam et al. [
For controlling network resources, Yanwen et al. [
RE-LEACH is a LEACH-based protocol proposed by Pandey et al. [
Miao et al. [
To extend the lifetime of WSN networks, Sahoo et al. [
Swarm intelligence-based forwarding (SIF) is proposed by Barzin et al. [
Sharmin et al. [
The proposed method provides a hybrid algorithm called over taker assisted political update (OA-PU) to identify the cluster head based on fusion-rider optimization algorithm (F-ROA) and political optimization algorithm. This system uses the k-means clustering technique to create the clusters within the sensor nodes. This technique improves energy consumption as well as network lifetime by considering CHs neighboring nodes. The primitive descriptions of the used F-ROA and PO optimization Algorithms are discussed first in this section for a better understanding of their benefits and limits.
In the typical WSN setting, there are multiple difficulties to be solved to enable better efficiency and energy optimization. The CH of a WSN is normally determined by factors such as energy, distance, cluster radius, and time. The parameters are taken into account as WSN is integrated with k-means clustering tools. The main goal of this effective CH selection using Fusion-ROA and PO is to minimize the distance between nodes, reduce execution time, as well as increase residual energy. The proposed CH selection model is a parametric representation. The batch size in a WSN often fluctuates, and CHs are chosen from the WSN’s total sensor nodes. When choosing a CH in a WSN, performance metrics such as node distance, latency, and spectral efficiency are taken into account, as previously indicated. As a consequence, only CHs may connect to the BS directly, and the access of all other nodes has been prohibited. When the characteristics from both WSN and IoT networks are integrated, the CH selection in WSN systems becomes more complicated. As a consequence, factors including energy, distance, cluster radius, and node time are used to calculate a low and mid-rate. This non-linear issue will be solved using our suggested OA-PU.
WSN comprises a group of stationary sensor nodes, each with its unique set of capabilities, designated by Sn. During data transfer, a node might operate as both a CH and an active sensor. Sensor location, topological features, data sensing, frequency hopping, and environmental impact are often connected to the WSN. The sensors are positioned at random or manually in application regions. Clustering is done by using a variety of sensor nodes. It’s a fantastic concept for prolonging the life of WSN. During the clustering phase, sensor nodes are gathered and a CH is chosen using the k-means clustering method, with the count denoted by Sc. This CHS is compatible with any cluster. In a cluster, the nodes are grouped such that they are as close to CH as feasible. The sensor nodes collect data on a particular region and send it to the CH during the operation. Furthermore, the collected data is sent to the BS by the particular CH. In the current study,
Selecting the correct cluster counts: In WSN, the cluster’s conceptual identity accounts are based on energy efficiency analysis. The use of a rational cluster count will improve network connection reliability while also extending network lifetime and balancing node energy loss. In inter-cluster communication, this architecture uses the multi-hop networking mode. Dj is determined by the interval between the BS and the severe CH, which is classified as several hops. Dj = k.
In
K-mean clustering: A k-means clustering algorithm is used to cluster the current results, which is described as follows: K-means is a classic data mining clustering algorithm that is often used to cluster massive amounts of data. This algorithm was first proposed by Mac Queen in 1967 and is known to be a simpler, non-supervised learning algorithm for resolving issues on recognized clusters. In a repeating method, the provided data items are categorized into k distinct clusters by resolving to premature convergence. As a result, the generated cluster’s outputs are separated and compressed.
This algorithm is made up of two distinct steps. The k pre-determined and fixed centers are selected at random in the first step. The stage after that delivers all data items to the nearest data center is referred to as the transport stage. The distance measure is typically used to compute the distance between each data object as well as the cluster head. The first stage is complete and an early collection is carried out, with all data items contained within a few clusters. The average of newly formed clusters has been recalculated in this manner. This procedure is iterated indefinitely before the minimal criterion function is achieved. The following are the measures involved in k-means clustering:
Input: k is the count of ideal clusters.
Output: k groups of clusters
Step: 1 from the specified nodes, choose k data objects at random as initial cluster heads.
Step: 2 Proceed;
Step: 3 Calculate the distance between each data object dbj(1 ≤ j ≤ n) and the entire cluster heads k Ci(1 ≤ i ≤ k), and allocate the data objects to the nearest cluster with dbj as mentioned
Step: 4 the cluster head is recalculated for each cluster i (1 ≤ i ≤ k).
Step: 5 Repeat when there is no change in the cluster head.
The competitive radius is used to describe the relationship between the cluster node count as well as the distance to BS
Following the creation of clusters, the best CH for data transmission must be chosen. This paper attempts to choose the best CHS by considering the fitness parameters such as energy, distance, cluster radius, as well as time, which have already been identified as issues when selecting the CH.
Since the shortest path selection improves the presentation of data transfer, the distance can be reduced. The use of energy by each node is another incredibly serious issue. As a result, the knowledge flow is the most important complexity with the least amount of energy and the shortest range. Since the clustering rate has increased over time, we need to determine the cluster radius. Almost all optimization algorithms place a high value on distance, cluster radius, and energy consumption when making a decision about which CH to use. As a result, various objectives must be met to extend the node’s mortality rate. Finally, while choosing the CH from a group of sensor nodes, energy, proximity, cluster field of view and duration must all be taken into account.
The Fusion rider optimization algorithm (F-ROA) is based on a unique computer approach known as fictional computing, which employs a series of processes to improve the system performance using fictitious facts and concepts. F-ROA relies on groups of riders that are attempting to accomplish the goal. To achieve this, F-ROA uses rider groups that go on a path to accomplish a shared goal. The number of groups in F-ROA is four, with equal riders allocated in each.
F-ROA is a novel computational algorithm based on the motivation of a group of riders who drive together towards a certain destination to achieve the cluster head position. They are divided into four classes, with the number of riders spread evenly among the four groups. Bypass rider, follower, over-taker, and intruder.
Riders create a multi-technique, with the proper formulation of the pedal, speed, guiding, and throttle being the most essential considerations while accomplishing the goal. Riders manage their position at each iteration by managing these characteristics and then start the well before strategy depending on the current energy rate, which is perpendicular to the direction between the rider’s goal and present location.
The current leading rider is listed based on the current achievement rate. This procedure will continue until the riders have been given the full amount of time. Following that, the cluster head is declared alongside the leading operator.
The optimal cluster counts k objective function OC1 is given in
In general, a WSN is made up of groups of clusters, each of which has its CH that acts as a coordinator for collecting data from sensor nodes and transmitting it to the BS. As a result, selecting the right CH, which has an impact on deciding the network’s overall system performance, is important. F-ROA is used to choose the best CH from a cluster of nodes. While F-ROA has many advantages for solving complex problems, it still needs to be improved to reach its full potential. In reality, the handling of community updates has been improved.
Rider’s position update: The pioneering rider is decided by upgrading the rider’s placement in each series, as shown in
Update procedure for bypass riders: Since they bypass the usual path and do not join the leading riders, the bypass rider’s location is updated at random. This is represented in
Update procedure for follower
Update procedure for overtaker: The relative progress rate, path predictor, and coordinate cluster head selector are all important factors in the overtaker upgrade process. The overtaker’s location change model is written in this manner
Update procedure for the intruder: When it attempts to take the leading rider position, the intruder uses the same mechanism as a follower to update its position. The method of updating the intruder’s location is depicted as follows
Each cycle t is compared to the preceding repetition t−1 to check if the current t produces an improvement. If the analysis is correct (getting the proper suggestion compared to the previous t−1), the process is carried out as usual. If not (the solution has not been enriched), the parameter trail is set to 0 and the iterations proceed. If the trail is greater than 5, the overtaker update is carried out as above
Energy: Energy consumption of non-CH nodes: After the CH has been selected; each non-CH node decides which CH will be unified in the new stage. The CH is chosen based on the estimation of the received transmitter capacity. As a result, the energy use up of a non-CH node updating the cluster can be calculated as follows:
The total energy required to create a cluster is calculated as follows:
Energy consumption of CH nodes
Distance: When data must be transferred from one node to the next, distance is a major factor. The distances are calculated using the wireless network, which is in control of the nodes’ mobility depending on their locations, velocity, as well as acceleration. The interval between the CH
To find the optimal result based on the fitness value, the suggested Political Optimization Algorithm (PO) is used. Each agent provides the solution in the proposed hybrid approach. For the fitness function, the optimal solution has the fewest set of data with the highest connection efficiency and kinetically selected CHs with the maximum residual energy. The foremost process is concerned with energy conservation and they recognize that reducing the optimal number of CHs will minimize energy consumption.
Senor nodes or various sensor nodes (i.e., parties) in a cluster are shown in
Even though the traditional PO model produces adequate results, it suffers from fitness function and is also prone to local optima that delay the optimization process. As a result, the PO definition is merged with the current WOA system, resulting in improved convergence as well as optimal outcomes. Certain search problems have been stated to benefit from hybrid optimization algorithms. The incredible hunting method of humpback whales is the most powerful motivating factor. They follow the predator’s movements and encircle it. The current optimal option, according to this technique, is the one that is closest to the target being sought.
It is important to note that J must be modified if the best outcome becomes available. The vectors C and E are enumerated according to
The search agent location is modified in the current hybrid system based on the PO update that is specified in
The hybrid scheme was created by combining the algorithms, all of which are innovative optimization approaches. In this scenario, both approaches concurrently offer the optimal best outcomes in each round of repetition; the best outcome is picked in each round of repetition by assessing the guidelines established by both refinements. Furthermore, the best solution is given to both algorithms to produce optimal solutions that are much better than the previous optimal solutions, as well as the cycle is repeated until the maximum number of iterations is reached. To solve a wide variety of optimization problems. There are several complex problems for which conventional methods are unable to have the optimal solution at any given time. The most recent meta-heuristic optimization techniques are F-ROA and PO. In the optimization approach, the following parameters are considered.
Cluster radius: In a WSN, clustering is the effective approach for extending the network lifespan. It requires forming clusters of sensor nodes as well as selecting cluster heads (CHs) for each cluster.
Measures the distance between nodes in the cluster and the cluster head. The square root of the cluster’s maximum position from some point.
Time: The hierarchical clustering solution for WSN and its round duration are quantified using the initial useable capacity of sensor nodes, as well as the aggregate energy consumed by CH and non-CH nodes in around. For the first round of the setup process, nodes are designated as CH.
The operations carried out in a round are split into two phases: setup and steady-state. In the setup process, clusters are formed, and in the steady-state phase, sensed data is transmitted to its expected destinations.
The OA-PU technique was used to apply the presented cluster-head selection model in the NS-3 device, and the results were observed. Energy, distance, cluster radius, and time were used to compare the performance of the adopted strategy to that of other existing techniques such as HSA-PSO, SDN-ECCKN, RE-LEACH, GWO, GA, SIF, and EH-WSN. The study was carried out in two experiments: one with 100 nodes and the other with 500 nodes, with the number of rounds, varying from 20, 30, 40, 50, 60, 70, and 180, etc. In a 100 × 100 m2 network topology, 50 sensor nodes are randomly distributed.
Three performance standards are included in the LEACH-M simulation experiments: network lifetime, energy consumption, and the number of data received at BS. The other two indexes, except for energy consumption, are assessed using the assessment criteria, the first node dies (FND), half of the nodes alive (HNA), and the last node dies (LND) (LND). LEACH-M receives more data at BS when compared to LEACH-C and EE-LEACH. When the area is increased to 300 × 300 m2, however, the number of data packets sent to BS decreases. This is because, in the WSN as a whole, fewer alive nodes are transmitting their sensory data to BS after some nodes die early. Simulation parameters are given below
Parameters | Value |
---|---|
Simulation area | 100 m × 100 m |
Number of nodes | 500 |
Transmission speed | Millisecond's |
Number of CHs | 10% of total nodes |
Data packet | 500 bytes |
Initial node energy | 5.5 J |
In the analysis, the proposed OA-PU system was compared to the HSA-PSO, SDN-ECCKN, RE-LEACH, GWO, GA, SIF, and EH-WSN schemes with various numbers of sensor nodes in the network.
The four network parameters were found to be superior with the proposed OA-PU on average with the compared systems since it can manage various sensor node characteristics that assist in cluster head selection. It’s also effective at maintaining the pace of utilization as well as the amount of exploitation suffered during the network’s collection of productive cluster heads.
The suggested OA-PU scheme facilitates the creation of a combination between the extraction and discovery rates introduced by the basic F-ROA as well as PO algorithms, which results in a higher percentage of sensor nodes. The number of lost nodes in the network visualized during deployment of the proposed OA-PU structure and the compared HSA-PSO, SDN-ECCKN, RE-LEACH, GWO, GA, SIF, and EH-WSN methods. Initially, the network’s number of sensor nodes is zero. When the number of rounds is raised from 0 to 500, however, the number of missing nodes continues to rise. While performing the planned OA-PU method, the proportion of lost nodes in the network is still decreased on average with the analyzed HSA-PSO, SDN-ECCKN, RE-LEACH, GWO, GA, SIF, and EH-WSN systems. The percentage of lost nodes in the network fully drains only after 480 cycles, whereas, the percentage of packet loss visualized during the execution of the analyzed HSA-PSO, SDN-ECCKN, RE-LEACH, GWO, GA, SIF, and EH-WSN S systems. The suggested OA-PU scheme achieves a significant reduction in the number of packet losses due to the addition of four parameters in balancing mechanisms, which extends the network lifetime.
In this research, a novel hybrid clustering method based on fusion –rider optimization algorithm (F-ROA) and political optimization algorithm (PO) known as OA-PU has been created with four stages: cluster construction, separation and merging, collecting CH, and data processing. In terms of energy, distance, cluster radius, and time metrics, the proposed model selects the best cluster head. Initially, the cluster of nodes is formed using used the k-mean clustering algorithm. A probabilistic model for selecting the appropriate number of clusters has been added to this. The purpose of this screening process was to separate as well as merge cluster structures based on experimental results, as well as to achieve the network’s non-uniform clustering layout. Eventually, the time lag optimized data energy storage system was employed in the spectrum sensing stage to decrease the information fusion effect on the node mobility. As a result, the proposed method reduced energy utilization since the execution time is less. Simultaneously, the reduced energy consumption leads to enhance the longevity of the network. Furthermore, the proposed OA-PU algorithm checks whether the collected data reached the appropriate BS. Also, maintains a high coverage ratio and eliminating redundancy. However, proper channel allocation is needed for a strong probability of effective reporting. Albeit, the proposed algorithm efficiently reduces energy utilization and enhances the lifespan of the network, energy storage efficiency and efficient reporting chance are not addressed because the focus of this article is limited to leftover energy that has not been harvested and that does not include self-interested successful production models. Therefore the future research will be based upon the limitations of this approach.
The authors with a deep sense of gratitude would thank the supervisor for his guidance and constant support rendered during this research.