Optimizing Hybrid Metaheuristic Algorithm with Cluster Head to Improve Performance Metrics on the IoT

The Internet of Things (IoT) has subsequently been applied to a variety of sectors, including smart grids, farming, weather prediction, power generation, wastewater treatment, and so on. So if the Internet of Things has enormous promise in a wide range of applications, there still are certain areas where it may be improved. Designers had focused our present research on reducing the energy consumption of devices in IoT networks, which will result in a longer network lifetime. The far more suitable Cluster Head (CH) throughout the IoT system is determined in this study to optimize energy consumption. Whale Optimization Algorithm (WOA) with Evolutionary Algorithm (EA) is indeed a mixed meta-heuristic algorithm used during the suggested study. Various quantifiable metrics, including the variety of adult nodes, workload, temperatures, remaining energy, and a target value, were utilized IoT network groups. The suggested method then is contrasted to several cutting-edge optimization techniques, including the Artificial Bee Colony method, Neural Network, Adapted Gravity Simulated annealing. The findings show that the suggested hybrid method outperforms conventional methods. the of grid and power consumption in machine-to-machine networks.

EA to improve utilization. The following is how the remainder is organized: Recent research on energy optimization of WSN devices in the IoT system is discussed in Section 2. Section 3 reviews the basics of the proposed system. Section 4 elaborates on the proposed framework. The results of the experiments are analyzed in Section 5. Results and other efforts are detailed in Section 6.

Materials and Methods
Many efforts have been made to extend the life of IoT networks. In this area, we've taken a look at some of the most well-known works. The authors of [9] proposed employing a metaheuristic method called quantum particle swarm to enhance the lifetime of the IoT systems to use a cooperative Multiple-Input-Multiple-Output strategy (MIMO). For every trip, QPSO seeks to automatically select the optimum collaborative numerous transmitter and receiver modules, leading to increased life. The researchers used additive white noise for similarity communications throughout this study.
The researchers employed three-hop networking in their experiments, and a simulation was run to test the channel's battery capacity for only 2 * 2 MIMO and 3 * 3 MIMO. The writers of Standard 10 suggest a neuro-fuzzy rules grouping method for improving edge efficiency. The method is designed to find the optimal CH and means knowledge effectively, lengthening the life of the network [10]. The calculation of CH energy consumption and CH to drain node range provides an overview of system performance. The simulation results show that the use of neuro-blur technology reduces power consumption while increasing the life of the grid. The researchers of this paper concluded that all nodes in the network are trustworthy.
The researchers of [11] provide a methodology for determining the optimal cluster prefix.
Because IoT devices have a shorter battery life, they have far less capacity and a shorter available bandwidth. The information obtained by the researchers was sent to IoT BS using this suggested data transmission. To enhance the network's durability, two of the finest clusters headers are picked in the same clusters during the CH selection procedure. The method of selecting dual CHs is an information entropy info fusion. The sensitivity information would then be utilized for categorization and fusing, data transfer, and categorization outcome blending [12].
The researchers of [13] use Self-Adaptive WOA (SAWOA) to create a power Channel access strategy for the WSN-IoT network. Latency, duration, energy, strain, and heat were all taken into account while determining the optimal CH efficiency. The suggested model is compared against several algorithms based on WOA, including ABC, CH selection methods, etc. The simulation model's findings extend the network's lifetime.
The authors of [14] utilize a unique MOFGSA method in choosing the optimal CH. The power of each node generates many messages for efficient transit through the IoT. The FGSA approach combines fractional and GSA theories and shows the performance of the method is compared to existing algorithms such as ABC, GSA and Multi-article Swarming Cooperative Method.
The researchers of [15] suggest a HEEQA method for achieving a balance among equipment power. The message authentication code underlay settings are then adjusted to reduce energy usage in the systems. In Connected systems, maintaining QoS is a big problem, and maintaining the carbon cycle is critical to extending the sensor's life. Quant particle swarm optimization (QPSO) is used with an improved framework's evolutionary sort technique to accomplish that.
The HEEQA technique improves energy consumption, thereby providing improved lifetime, speed, and supplies, according to the simulated results.
The authors of [16] suggest an unequal cluster of the period routing algorithm to solve the problem of unbalanced energy usage and delayed data transmission in WSNs. It is contrasted to several other algorithms. Simulation trials revealed that such an approach enhances system performance and balances the channel's energy usage, increasing the channel's life. The proposed method is especially well-suited to IoT systems that need low latency. Describes a data sources correlation algorithm that increases cognitive network bandwidth by combining data correlation analysis with device grouping [17].
The writers of [18] proposed a general IoT Data Streams cluster-based approach that uses the distribution of data to discover different categories identified in the stream of data. The suggested grouping technology's findings may be utilized as input for information processing, incident acknowledgment, and actual data stream analytics.
The researchers of [19] propose a new solution for edge devices with Connected systems called non-orthogonal multiple access (NOMA), which takes into account IoT device clusters and energy management in Hybridized NOMA systems. The complex clustering of IoT devices is addressed in the early phases to reduce network complexity and offset of improved channel conditions for IoT devices. In the second phase, Nash's balanced approach is used to address energy management based on electricity distribution. The simulation indicates that the proposed strategy can improve spectrum efficiency while interacting with other IoT systems and devices.
A Data Cluster Systems for Privacy and Availability (PADC), which uses k-mean methods and differentiated data protection through one central location to the other. Every iteration, the PADC algorithm multiplies each clustered value by the frequency of clusters, assisting in the calculation of the different positions among pieces of data for more precise input numbers. When compared to previous differentiated methods in terms of security, the suggested scheme improves the availability of sufficient bandwidth while maintaining the same level of privacy, implying that the suggested PADC scheme surpasses previous smart IoT electric grid solutions [20].
Based on the foregoing explanation, this can be stated that, although there are so many CH selection systems available, they all consume a great deal of energy. To solve this problem, a hybrid WOA-SA method is applied to minimize power consumption in the IoT network by selecting the best CH.

IoT WSN Adaptive CH Control
Many sensor networks make up an IoT network, each with limited storage and high energy consumption [21][22][23]. Because these units are constantly generating data, the batteries should be used more often. The lifespan of a system is shortened when energy usage is higher. One of the strategies of energy optimization is to select the best CH. Clustering is a process for dividing sensors into subgroups and assigning a leader from each grouping given a set of criteria. Cluster is known as subgroups, while CHs were known as facilitators. Cluster aids in the efficient use of internet services, network congestion control, packet forwarding, energy consumption, and life of the network management. Transmission cost inside the clusters, sensing power, and node distance from the BS are among some of the characteristics utilized to choose the optimal CH.
The cluster nodes will transmit the information to their CH following picking the best CH. The data gathered would be sent to centralize BS via CH.
Only the TCH that has been picked can connect with the IoT core satiety. Selecting the wrong CH for an Internet -of -things WSN which results in a life improvement became a difficult challenge.

Fitness function
CH is selected in classic WSN depending on characteristics such as distance, latency, and energy. We must focus on additional factors like power & temperatures even though we are combining WSN with IoT [24]. As a result, to optimize network performance and net longevity, the CH is chosen based on even a cluster having strong power and high load, as well as latency, length, and heat. The fitness value, which would be represented in Equation (1), must be maximized to increase the channel's stability and reliability.
In the subcategories that follow, the statistically modeled estimate of the five variables established in this study is discussed.

Calculation of power
In IoT systems, the power spent in a single channel may be divided into two parts. The very first element is the power used by all Power Amplifiers (PA), and the other is the power used by the other Circuit Blocks (CB). For each connection, the power consumption, E, is calculated by using the following: IoT node energy can indeed be replenished. The power of an IoT node was originally E0. The packet must be transferred to a CH by each node in the network. Both CH, as well as the node, lose energy when transferring messages from the x th node to the y th CH. The hardware for the transmitter and receiver is present in each IoT device. The power evaporates in the form of a transmission and reception when the user transmits or transmits information. The transmitter's heat transfer is a type of energy amplifier known as radio circuits, whereas the recipient's heat transfer is also caused by radio electronics.
Power can be dissipated in 2 directions. The first one is when the user transmits m byte of memory to the CH specified in Equations (3). Whenever the CH gets m tb of memory from the base station, the heat transfer is shown in Calculation (5).  (7). The information is sent to the CH by the normal node till the node's power isn't negative. If the energy of a node falls to zero, it is referred to as a dead node. The equation expresses the objective functions for energy (8). Another of the performance characteristics used to identify the best CH is efficiency [25]. To be selected as CH, the node's energy must be sufficient.
The performance index of the range from sink node to IoT BS is theoretically demonstrated by range calculation Equation (9). To select the optimal CH, the range among clusters and the IoT BS must be kept under control.

||
− || and || − || are distance between normal mode with CH and CH with IoT respectively. M and N denote a range of dimensions in meters.

Calculation of the latency
The latency must be as little as possible while choosing the optimal CH. The delay will be between 0 and 1 milliseconds. Because the latency is determined by the number of nodes in the cluster, the number of nodes in a cluster must be kept to a minimum to reduce the delay. The equation shows the mathematical fitness function of the delayed transfer between Connected systems and CH [26]. The fraction is the highest transmission from a CH to the BS, while the denominator is the maximum set of neurons.

The WOA algorithm
WOA is a heuristic approach that is influenced by the environment [27]. This technique was designed depending on humpback commercial whaling habits. The bubble-net feeding technique is a whale assault style in which bubbles are generated in a ring from around the victim. There are two stages to this process. Attacking Prey, encircling Prey. Whales have a distinctive foraging habit that employs two strategies. One would be the uptrend, in which whales dive 12 meters below the surface and begin generating spirals of bubbles around their prey. Loop catch, natural selection to increase, and looped corals are the 3 parts of the second method.

Surrounding the prey
The humpback attempts to locate its meal and afterward surrounds it. The whales think that the present best outcome is target food so alert other searching whales to shift their positions towards the target food. Equations (11) and (12) depict the surrounding prey process [28].

The assault of the bubble-net
There are two ways to assault a whale shark using a bubble net. The decreasing circumferential method and spiraling update are the first two approaches. The humpback employs a few of these two tactics to attack its food. The chances of catching prey using these methods are about 50%.
Approach with a decreasing circumferential effect. Now, the quantity of "x" begins to decrease from 2 to 0, affecting the value of X in Equation (13). X is a special variable in the [x, x] range. Figure 3 shows the variety of integer variables for X in the region of [1,1].
Helix upgrading is a method of updating that uses a spiral. The equation is used to compute the distance between the whale and the target food in this method [29].

Hunting for prey (exploration phase)
During this stage, the whales look for the best solution (prey) throughout the globe at random and shift their status in comparison to other whales. The value of X must be larger than 1 to compel its search process to go far away from several references whale worldwide. Equations (17) and (18)

Simulated annealing
Simulated Annealing (SA) is also an optimization process based on a thermodynamics notion, namely how metals cool and then anneals. The 34 SA algorithm is used to choose the leading international optima from a set of local maxima. The optimization problem of a related optimal solution is used in this approach [31]. The road approach is comparable to this optimal solution.
The optimum step for both the goal function will be picked in the hill-climbing approach, but a randomized movement would be selected in EA.
This will be approved if the selected randomized movement enhances the result. Otherwise, the method computes a decision with a chance of less than one. Figure 4 shows the architecture of the proposed WSO-SA method for the CH selection process in an IoT network. A BS with clustering makes up the created virtual IoT network. Several endpoints, as well as a CH, would be displayed for each cluster. The information is sent to the CH from each node. The CH will compile the information gathered from the entire node and transmit it to the IoT BS. The combination WOA-SA method is utilized to find the optimal node for CH. A variety of loading, cost, and thermal operating conditions was evaluated while selecting the CH. The following is the code for the proposed framework.  For all techniques, the temperature mostly on IoT devices appears to be lower at first. The temperature competitive advantage arises as to the number of cycles increases for all techniques. Figure 5 shows that the temperature produced by all of the techniques is almost identical at first. However, after 1800 cycles, the heat created either by the suggested technique is significantly lower than that produced by earlier studies.

Evaluation of the effectiveness depending on the load measurement
The effectiveness of the suggested system that is based mostly on the loading of CHs is shown in Figure 6. The efficiency of the IoT system would be best if the load is evenly spread among several CHs. It can be seen from this diagram that the suggested model evenly spreads the load among all CHs. In all rounds, the suggested model optimizes the workload. As compared to existing algorithms, the burden on CHs is negligible even after 1500 cycles. The temperature generated by the sensor network would be lower whenever the strain upon this CH is lower. The IoT system consumes lesser energy with the help of all this and enhances the program's efficiency.

The cost function is used to assess project
The effectiveness of such a proposed approach based on a cost function is shown in Figure   8. The efficiency of a method is determined by the cost function. The efficiency of methods will improve significantly as the number of samples rises. The proposed approach exceeds the current methods in terms of converging, as shown in the graph.

Figure 8: Cost Function Performance
In almost all of the metrics evaluated, the current platform outperforms the traditional approaches, as shown in Figures 5-8. Linear approaches, regardless of performance parameters, employ the blinded operators for exploitative reasons. The proposed algorithm, which works as an operation throughout the WOA-SA technique, guarantees that the blinded operators are substituted with such a search engine that uses the result as an origin point. Eventually, following focusing on one aspect searching, the improved result fills the role of the original. As a result, the WOA application's exploiting capability is improved by the evolutionary programming process. As a result, the modeled aging technique aids in the WOA's effectiveness in locating the best solutions. As a result, the proposed method outperforms current methods in terms of IoT system performance optimization.

Conclusions
Even though IoT has enormous opportunities in a spectrum of uses in the modern-day, there are several obstacles to overcome. Safety, power efficiency, transportation, hardware configuration difficulties, information network connectivity, and other troubles must be solved to improve the resilience of IoT. Researchers opted to focus on the optimized design problem in this study. To solve this problem, the current study employs a hybrid metaheuristic method called WOA-SA to optimize the power usage of devices in IoT-based WSNs. The Xively IoT technology is utilized in this study to mimic an IoT network. The Internet of Things connection has gone through 2000 versions. Throughout this study, various performance measures including load, power consumption, number of living nodes, cost function, and temperatures were utilized to select the best CHs for IoT systems. The suggested method is then compared to several existing methods. The results of the experiments show that the suggested method outperforms the traditional methods. Numerous other performance criteria, including latency, node density, link lifespan, and so forth, might be evaluated throughout the latter to pick the best CH. In addition, the suggested work may be scaled by utilizing this technique in real-time systems with a large number of sensor nodes, and the energy usage of the whole IoT system is assessed in this study. Future studies may look at power management in each isolated network in an IoT system, and safety issues identified by multiple researchers like reference 33 could be used in the proposed model to solve security problems.