TELK OMNIKA T elecommunication, Computing, Electr onics and Contr ol V ol. 23, No. 6, December 2025, pp. 1718 ∼ 1728 ISSN: 1693-6930, DOI: 10.12928/TELK OMNIKA.v23i6.27007 ❒ 1718 Escalating QoS by r ey optimization of CGSTEB r outing pr otocol with subordinate ener gy alert gateways R. Madonna Arieth 1 , Ramya Go vindaraj 2 , Subrata Cho wdhury 3 , Thi Thu Nguy en 4 , Duc-T an T ran 5 1 Department of CSE, V el T ech Rang arajan Dr .Sagunthala R&D Institute of Science and T echnology , Chennai, India 2 School of Information T echnology and Engineering (SITE), V ellore Institute of T echnology (VIT), V ellore campus, T amilnau, India 3 Sri V enkatesw ara Colle ge of Engineering and T echnology (A), Andhra Pradesh, India 4 School of Electrical and Electronic Engineering, Hanoi Uni v ersity of Industry , Hanoi, V ietnam 5 F aculty of Electrical and Electronic Engineering, Phenikaa Uni v ersity , Hanoi, V ietnam Article Inf o Article history: Recei v ed Feb 22, 2025 Re vised Sep 23, 2025 Accepted Oct 19, 2025 K eyw ords: Clustering Firey optimization General self-or g anized tree-based ener gy balancing Routing protocols Subordinate ener gy alert g ate w ays W ireless sensor netw orks ABSTRA CT W ireless sensor netw orks (WSNs) comprise lar ge numbers of sensor nodes that are highly constrained by limited battery po wer , making ener gy-ef cient routing essential for sustaining netw ork lifetime and service quality . A mong e xisting solutions, the general self-or g anized tree-based ener gy balancing (GSTEB) pro- tocol with clustering has been widely adopted for ener gy-a w are communication. Ho we v er , GSTEB and its clustered v ariant often suf fer from ener gy imbalance, high pack et loss, and reduced quali ty of service (QoS) due to e xcessi v e load on cluster heads (CHs). T o address these challenges, this paper introduces an enhanced routing frame w ork that inte grates rey optimization with clustered GSTEB (CGSTEB) and introduces subordinat e ener gy alert g ate w ays (SEA Gs). The rey algorithm is applied to optimize CH selection through a tness func- tion that balances residual ener gy and node proximity , ensuring ef cient cluster formation and adapti v e load distrib ution. Meanwhile, SEA Gs establish a tw o- hop communication model between CHs and the base station (BS), reducing CH ener gy consumption and pre v enting premature node f ailures. Simulation e xper - iments conducted in NS2 demonstra te that the proposed rey-CGSTEB with SEA G signicantly impro v es QoS metrics, including netw ork lifetime, ener gy utilization, throughput, and pack et loss rate, compared with con v entional CG- STEB. T hese results conrm the ef fecti v eness of combining metaheuristic opti- mization with g ate w ay-assisted routing for resilient and ener gy-ef cient WSNs. This is an open access article under the CC BY -SA license . Corresponding A uthor: Thi Thu Nguyen School of Electrical and Electronic Engineering, Hanoi Uni v ersity of Industry Hanoi, V ietnam Email: thunt@haui.edu.vn 1. INTR ODUCTION W ireless sensor netw ork (WSN) has been utilized in di v erse elds lik e the military , health care ser - vices, smart city , remote monitoring, industrial automat ion, and agriculture. In healthcare, sensor operation guarantees uninterrupted patient monitoring. In smart cities there is lo w-ener gy communication in traf c, en vironmental, and utility monitoring. Industrial automation and precision agriculture can le v erage this proto- col to maintain scalable, ener gy-sustainable sensor deplo yments, minimizing do wntime and operational costs. J ournal homepage: http://journal.uad.ac.id/inde x.php/TELK OMNIKA Evaluation Warning : The document was created with Spire.PDF for Python.
TELK OMNIKA T elecommun Comput El Control ❒ 1719 Furthermore, in critical domains such as military surv eillance and disaster management, where real-time and ener gy-a w are communication is vital, The netw ork consists of man y sensor nodes that closely monitor the ph ysical and en vironmental conditions [1]. A sensor node collects, aggre g ates, and sends the en vironmen- tal and ph ysical information to base st ation (BS). Due to t he nite node batteries, the netw ork’ s lifespan will quickly deplete whene v er data is transmitted [2], [3]. In most of the current protocols, mult i-hop communi- cation is assumed. It might generate ener gy hole problems, especially when sensor nodes (SNs) are close to the BS, including unnecessary transmission is o v erhead [4]. Numerous clustering protocols are proposed to a v oid ener gy-hole problems and e xtend netw ork lifespan [5], [6]. Some protocols based on ener gy-ef cient routing played a signicant role in enhancing the WSNs’ operations [7], [8]. There are lot of routing protocol such as po wer -ef cient g athering in sensor information sys tems (PEGASIS) is a chain based routing protocol, lo w-ener gy adapti v e clustering hierarch y (LEA CH) general self-or g anized tree-based ener gy-balance (GSTEB) is an tree based ener gy ef cient routing protocols. The ener gy ef cient clustering systems for WSNs is di vided into tw o forms. The clustering systems enforced in homogeneous WSNs are called homogeneous clustering systems, and another one which is enforced in heterogeneous WSNs are called heterogeneous clustering sys- tems.Here heterogeneous clustering system is used. In clustering techniques in WSNs focus on collecting data within groups of nodes, where leaders are elected from among them. These leaders, or cluster heads, are re- sponsible for aggre g ating the data and transmitting it to the BS. I n clustering, the netw ork is di vided into small groups. Each group chooses a single node as the cluster head (CH), and the e v erlasting will be a non-cluster node [9], [10]. Thus, clustering is a signicant perception for consistently spreading ener gy con v ention and e xtending netw ork lifetime in the WSN [11]. The elect ion of the CHs plays a signicant role in reducing the netw ork consumed ener gy and their distrib ution o v erall in the monitoring area. F or instance, the election of cluster heads proposed by the GSTEB protocol, [12], [13] is based on randomness, and cluster headcount uc- tuates greatly . Figure 1 sho ws one arrangement of GSTEB clustering with a single hop to the base station [9], [14]. But GSTEB pro vides a strong baseline for ener gy-a w are routing b ut f ails to address multi-objecti v e opti- mization, load balancing, and adaptability T o enhance, the opt imization techniques are utilized in much of the literature [5], [7]. These techniques focus on optimized routing in WSNs such as particle sw arm optimization (PSO) ant colon y optimization (A CO), and man y others. But, one of the proposed protocols is the clustered GSTEB (CGSTEB) with optimization techniques [4]. As the name suggests, it forms a cluster based on a tree structure. The algorithms’ main challenge is to elect cluster heads with the highest ener gy and close enough to other nodes to reduce the data communication distance [14]. So, PSO optimization, for instance, tries to achie v e a node with the best global and particle best v alues. The global best indicates the current data particle close to the objecti v e particle, while the particle best represents the contiguous molecule information that has e v er gone to the objecti v e. The main dra wback of PSO optimization is that it is easy to f all into local optimum into high computational space, which af fects the qualit y of service (QoS) [15], [16]. Our proposed w ork tries to o v ercome the dra wbacks of PSO and in v estig ates the ef fecti v eness of applying rey optimization [17], [18]. Firey optimization introduces a more intelligent and adapti v e mechanism for CH selection and routing. The main idea of the optimal rey approach is to k eep a w ay from the local minimum problem, and it will also transmit the data e v en though it is v ery noisy . The rey heuristic is based on the light intensity produced by reies. The intensity of light produced is mapped to the objecti v e function; hence, reies with lo w i ntensity are attracted to reies with higher light intensity . A h ybrid v ersion of the of the rey algorithm, the synchronous rey algorithm, is proposed based on the insect reies, which will produce light. Here insect reies, where v er a brighter sensor node in terms of ener gy and distance, will attract the less bright neighboring sensor nodes. Thus, the less bright sensor node can depend on the brighter sensor node for data transfer , sa ving ener gy . A tness function has been designed based on the combination of tw o parameters, ener gy and distance, which decide the brightness of the sensor node [19]. But only at the lo w e xploration capability of rey , which is al w ays in one direction, making it impossible at times to achie v e optimal solutions. This can be solv ed by using comple x problems, an y function, or fractional order here. Because in the meta -heuristic rey algorithm, randomly generated solutions will be considered as reies, and brightness is assigned depending on their performance on the objecti v e function. When it mo v es for con v er gence, the algorithm is indirectly proportional to the number of reies. so it will run for se v eral times till it reaches the con v er gent parameter set. If dif ferent results are obtained, it will be too high; it reaches non-optimal points. A slo w con v er gent parameter set should be used, i.e., a lar ger number of reies and a greater domain size. If t he same result is obtained ag ain and ag ain, b ut during a lar ge fraction of the algorithm. Our proposed w ork o v ercome ener gy imbalance, high pack et loss, and reduced QoS issues in traditional clustering-based protocols by intelligently selecting cluster heads Escalating QoS by r ey optimization of CGSTEB r outing pr otocol with ...(R. Madonna Arieth) Evaluation Warning : The document was created with Spire.PDF for Python.
1720 ❒ ISSN: 1693-6930 and g ate w ays. The ne xt phase is data transmission, where data is transmitted to the base station through the CHs, which produces a hea vy b urden on CHs. Therefore, CHs ener gy might get depleted quickly . One of the solutions is proposed as in [3], it i s a ne w subordinate ener gy alert g ate w ay (SEA G) to be implemented, as sho wn in Figure 2. The g ate w ay load is minimized when the load balancing scheme is allotted to sensor nodes. So, in this paper , the g ate w ay technique is applied for data transmission. The reusable g ate w ay node is inserted into the netw ork since the g ate w ay node is rechar ged based on the location details [20], [21]. Our main objecti v e is to reduce the ener gy and increase the life span of the netw ork. When it i s deplo yed in a remote area, when data is transferred from source to destination, ener gy will be e xhausted. T o a v oid this, our proposed w ork is done. The primary objecti v e of inte grating rey optimization with SEA Gs g ate w ays for f ast con v er gence , and also frame w ork is designed to balance multiple Qo S objecti v es such as ener gy ef cienc y , f ault tolerance, and end-to-end performance ultimately enabling self-optimizing and resilient netw ork operation. T o a v oid the depleted ener gy of CHs, the y are e xchanged in rounds where ne w CHs are chosen in each round. The CH is elected autonomously depending upon the position and distance of the g ate w ay and base station. The nodes send t heir data to CHs, and CHs aggre g ate the recei v ed data and send the aggre g ated data to the base station. The cost of the g ate w ay is less when contrasted with dif ferent nodes [21], [22]. Figure 1. Clustering of sensor nodes Figure 2. SEA G node with CGSTEB 2. RELA TED W ORK The follo wing part discusses the e xisting WSN, ener gy-ef cient routing protocols, and opt imization techniques utilized for such protocols. Se v eral ener gy-ef cient routing techniques encompass to increase en- er gy , b ut satisf actory results were not found. F or instance, Liu et al. [1] proposed an e x ecuti v e po wer scheme necessary for rotating inedible se gments to mak e it possible with a specic time. Similarl y , impro v ed ener gy- ef cient (IEE)-LEA CH [2] is a clustering algorithm proposed to reduce ener gy consumption and prolong the lifespan of sensors. The limit of the proposed IEE-LEA CH con v ention had se v eral boundaries lik e initial, resid- ual, total, and a v erage ener gy so that it can pick up the system’ s po wer . Deepa and Rekha [23], the GSTEB w as proposed in the direction to construct routing tree techniques e v erywhere in support of e v ery single round, the BS assigned a root node and communicated to e v eryone. After that, e v ery node elects its head by con- sidering only the neighbor’ s information, making GSTEB a strong con v ention. Han et al. [17] e v aluated the TELK OMNIKA T elecommun Comput El Control, V ol. 23, No. 6, December 2025: 1718–1728 Evaluation Warning : The document was created with Spire.PDF for Python.
TELK OMNIKA T elecommun Comput El Control ❒ 1721 GSTEB in terms of netw ork lifespan and load balancing. Shah et al. [19] tried to de v elop a multi-hop cluster routing protocol, multihop-LEA CH, that follo ws cluster intra and inter -multi-hoping. Sujae and Arulselvi [9] proposed in WSN there will be ef fecti v e data o w among the nodes in the clusters. Here ener gy-ef cient mod- ied clustering (EMoC) algorithm w as proposed for e x ecutes the clustering and cluster head selection. W ang et al. [20] de v eloped there are other techniques that use group pattern and CH selection by PSO optimization, genetic algorithm has been e xpansi v ely depending on CH select ion techniques through multi intention utility taking into account po wer utilization along with pack et delay . In 2020, sho ws that when we are using a triax- ial (MEMS) accelerometer it is a lo w-cost sensor f which is used f o r measuring the acceleration. The v alues measured by the sensor are noisy and inaccurate. Therefore, calibration algorithms needs. In the case of an accelerometer , using the magnitude of the gra vity v ector as a stable reference leads to a nonlinear optimization problem. It is achie v ed by cuck oo optimi zation. But the dra wback of nonlinearly optimization will increase the number of iterations. Ahad et al. [21] re vie w the rey algorithm’ s usage in v arious rele v ant domains. The authors accompli shed that it can control multi-modal problems and ha v e quick con v er gence and a lso conned look for heuristic. The authors of [24], [25] proposed that half con v ention lo w ener gy-a w are g ate w ay (LEA G). Gate w ay is utilize d to limit the vital ity utilization, and information is sent to the base station. It used Zigbee to diminish ener gy utilization and routing. Mehta and P al [26] proposed an optimization technique, which is the formulation of determining the best solution for purposeful as well as v aluable as possible for limiting or augmenting the parameters. PSO optimization easily f alls into local optimum into high computational space so that it af fects the QoS. 3. PR OBLEM FORMULA TION Here, it is assumed that the sensor netw ork is represented in the form of graph G = ( V , E ) with a set of v ertices V = { V 1 , V 2 , . . . , V n } and edges E = { E 1 , E 2 , . . . , E m } . It is als o assumed that each edge has a weight, see (1). W i = ( W 1 i , W 2 i , W 3 i , . . . , W pi ) i = 1 , 2 , . . . , m (1) where W pi , p = 1 , 2 ... is the weight of each edge p. x = x 1 , x 2 , . . . , x n is identied using (2): with x being the current location: x i = 1 0 if E i sel ecte d other w ise (2) The objecti v es were formed as: z 1 ( x ) = X n i =1 w 1 x i (3) The objecti v es were formed as: z 2 ( x ) = X n i =1 w 2 x i (4) M inimum z p ( x ) = X n i =1 w n x i (5) where, Z p ( x ) is the i th objecti v e to be minimized for the problem, where w 1 , w 2 , . . . w n are weighing parameters (normalized v alues), C denotes current node, i is the member and n denotes the number of members co v ered within the cluster . Here are three natures of administrati v e boundaries, pack et loss rate, and one-w ay delay also remaining vitality is considered to manuf acture the tar get w ork as a minimization. 3.1. Experimental The e xperimentation contains v arious steps essential to carry out the intention: Step 1: Setting up the netw ork; Step 2: Deplo ying sensors; Step 3: Creating SEA G g ate w ay node; Step 4: Creating SEA G g ate w ay node; Escalating QoS by r ey optimization of CGSTEB r outing pr otocol with ...(R. Madonna Arieth) Evaluation Warning : The document was created with Spire.PDF for Python.
1722 ❒ ISSN: 1693-6930 Step 5: Clustering GSTEB appropriately to estimate le v els; Step 6: Selecting CH and formation by utilizing the rey optimization technique; Step 7: Computing the sensor’ s ener gy le v el; Step 8: Stopping. 3.2. GSTEB clustering algorithm GSTEB is a tree-based routing protocol that uses clustering to group nodes and select cluster heads. Initially , nodes are grouped together to form clusters. The cluster operates in a circular round-based manner . F or the formation of cluster GSTEB, one node will be elected, and it is compared with the threshold v alue T ( n ) , then CH nominated moreo v er it residue as a habitual node. The threshold is gi v en as (6), where [*] is dened as a mathematical multiplication. T ( n ) = p 1 − p ∗ r mo d 1 p if n ∈ G (6) 3.3. Pr oposed framew ork f or r ey optimization The proposed algorithm le v erages the rey optimization te chnique to achie v e ef cient CH selec- tion and ener gy-a w are routing in WSNs. Firey is a meta-heuristics algorithm that functions on the ordinary ick ering light of reie where brightness and attraction guide their mo v ement to w ard optimal solutions. By modeling light intensity as tness v alues and updating positions iterati v ely , the algorit hm ensures minimized ener gy consumption, balanced load distrib ution, and enhanced netw ork lifetime. The elements of reies are as per the follo wing: (i) Ev ery rey can be eng aged to another independent of the same gender . (ii) The rey’ s sparkle is compared with its attraction, and among pairs, which increasingly shines that will attract the one with minor brightness. A rey will mo v e randomly if it can’ t nd an y increasingly splendid neighboring reies. (iii) Then, the scientic reproduction, the rey’ s brilliance, depends upon goal w ork. Mainly , it is a meta-heuristic capable of pro viding the nest solution to solv e a multi-objecti v e prob- lem. By using the ne w tness function, quality of services lik e ener gy , loss of pack et, and o v erall delay from source to destination are gi v en by: F ( x ) = ( P dr / P tot ) × ( E r / E in ) e − e dl / e mx (7) where, P dr presents the number of pack ets dropped; P tot indicates the total number of pack ets sent; E r is the residual ener gy in nod e i ; E in is the initial ener gy . e dl is the end-to-end delay; e mx is the maximum allo w able delay . The cluster formation and CH selection in reies are gi v en. In the rey calculation (Mehta et al. 2017), the v ariety of li g ht force and the issue’ s plan for eng aging quality are ur gent as the tar get w ork is encoded into it. The light po wer is determined to utilize γ ; the x ed light ingestion coef cient and the light force I can be gured dependent on separation r with the end goal that: I = I 0 e − γ r (8) where, I 0 is the beam strength. The appeal β of a rey is yielded. β ( r ) = β e − γ r 2 (9) where v er β is the charm at r = 0 . The space connecting pair nodes of reies be able to calculate by their Euclidean distance as: r ij = q ( x i − x j ) 2 + ( y i − y j ) 2 (10) A rey i mo v es to a more attracti v e rey j by: x i = x i + β e − λr 2 ( x j − x i ) + α (rand − 1 / 2) (11) In the proposed, it is cate gorized then the greatest one is elected as challenges. The elect ed one wi ll replicate themselv es by crosso v er and transformation. The ne w-f angled one is inserted into the group, and subsequently , iteration is sustained. Ef cient data transmission is an important aspect of WSN. T o transmit the data from BS will reduce the ener gy , so here we propose a SEA G to increase the ener gy between BS and g ate w ay . TELK OMNIKA T elecommun Comput El Control, V ol. 23, No. 6, December 2025: 1718–1728 Evaluation Warning : The document was created with Spire.PDF for Python.
TELK OMNIKA T elecommun Comput El Control ❒ 1723 3.4. SEA G T o reduce the b urden of the cluster head, A ne w no v el-based SEA G is proposed so the cluster head close to the g ate w ay will transmit the message and then transmit it to the sink. When the g ate w ay node po wer is lo w , it can be rechar geable so that its cost is reduced. The time di vision multiple access (TDMA) time scheduled is calculated for each round. The distance is calculated between the g ate w ay and the base station. The Zigbee is used to connect the nodes to the g ate w ay . It also sa v es ener gy and reduces the cost of the netw orks. So the transmission and process should be good quality . As sho wn in Algorithm 1 and Algorithm 2, the proposed SEA G method inte grates the rey optimization and g ate w ay-selection processes to enhance ener gy ef cienc y . Algorithm 1 Firey optimization technique Input : Declare x = x 1 , x 2 , .... x m Let s be the primary population x i , ( i = 1 , 2 , ...n ) Output : The beam strength I s on x i be intent on Pr ocess : Describe light inclusion coef cient γ while ( s < Maximum) f or i = 1 to n f or j = 1 to i if then ( W i > W j ) { The mo v ement of light emulation is between i through to j } end if { Charm di v er ges amid the distance r via e − γ r ; estimate ne w-f angled solution and bring up to date brightness strength } end f or j end f or i Ranking rey along with getting the recent nest; end while Algorithm 2 SEA G g ate w ay Input : Distance and the place of the g ate w ay node of coordinate ( x, y ) ; Let ( n ) be the threshold v alue; Output : The result indicates that the space among g ate w ay node and base station is calculated for e xpenditure reduction. Pr ocess : Begin Sends an advisement message to the g ate w ay node; { F or e v ery CH and g ate w ay node, If (CH == x && y) { locate x = x 1 and If g ate w ay node turns = f alse The g ate w ay node with locations x and y is indicated as x 2 and y 2 ; { If it is a g ate w ay node the space and the position are calculated as: G ( i, j ) = s ( i ) E s ( i ) ma x + d ( i, j ) x 2 + d ( j , x ) x 2 d ( j , s ) d ( i, s ) (12) } else Euclidean-distance between CH and g ate w ay node is a nearer distance; If nearer space < minutest distance Current distances is no w assigned to minutest distance Cluster g ate w ay node ID is assigned as a nearer g ate w ay node to cluster -head; else g ate w ay node turns = true; } End 3.5. Ener gy model In WSN, a vitality model is intended to compute ener gy loss in e v ery sensor node while commu- nicating with other sensors. There are tw o types of communication channels: T w o-w ay channel distrib ution and multipath channel for pack et transmission by means of multi-jump are utilized here. Hence, the ener gy e xhausted for transmission of pack et o v er distance is determined by: E t x ( k , d ) = E el eck ∗ k + e f s ∗ k ∗ d 2 , if ( d < d 0 ) E el eck ∗ k + e amp ∗ k ∗ d 4 , if ( d > d 0 ) (13) Escalating QoS by r ey optimization of CGSTEB r outing pr otocol with ...(R. Madonna Arieth) Evaluation Warning : The document was created with Spire.PDF for Python.
1724 ❒ ISSN: 1693-6930 where, e f s is free space ener gy loss, e amp is a multipath loss, d is a distance between source and destination nodes, and d 0 is crosso v er distance: d 0 = r ε f s ε amp (14) The ener gy spend by radio E RX ( k ) = k ∗ E elect (15) 4. RESUL TS AND DISCUSSION The proposed rey-CGSTEB with SEA G frame w ork w as e v aluated using the NS2 simulator . The netw ork w as deplo yed o v er a 100 × 100 m² area with the base station x ed at coordinates (100,100). A t otal of 300 sensor nodes were randomly distrib uted to emulate realistic deplo yment. Each node w as assigned an initial ener gy b udget of 0.01 J. Communication parameters were set to a transmission po wer of 60 nJ/bit, reception po wer of 60 nJ/bit, and amplier ener gy of 100 pJ. T o ensure statistical reliability , each e xperiment w as repeated 30 independent runs, and the reported results represent the mean v alues with 95 % condence interv als. Standard de viations are pro vided in the performance tables, and statistical signicance w as conrmed using t-tests comparing the proposed approach with baseline protocols (con v entional CGSTEB and Firey-CGSTEB). T able 1 sho ws a range of the simulation parameters used by the NS2 simulator . These parameters, such as pack et loss rate, one-w ay delay , ener gy consumption, and throughput, are sho wn as in Figures 3 and 4. T able 1. General simulation parameters P arameters V alues Area of the netw ork 100,100 Position of base station 100,100 T otal number of nodes 300 Battery early po wer 0.01 Battery transmitter po wer 60 nJ/bit Battery recei v er po wer 60 nJ/bit T ransmit amplier 100 pJ Ener gy for aggre g ation 5 nJ Max lifespan 200 Message range 2000 bits Figure 3. P ack et loss rate Figure 4. End-to-end delay In the pack et loss rate, the number of nodes is v aried from 50, 100, 150, 200, and 250; with the pack et loss, where the node increases, the pack et loss rate is decreased. It denotes that the proposed rey CGSTEB with SEA G g ate w ay is more ef cient when compared to the e xisting g ate w ay . The pack et loss rate of the proposed rey CGSTEB with SEA G has 86% of lesser tha n the pre vious Firey based CGSTEB. In Figure 4, end-to-end de lay is the amount of occasion in use to broadcast a frame through the system from be ginning TELK OMNIKA T elecommun Comput El Control, V ol. 23, No. 6, December 2025: 1718–1728 Evaluation Warning : The document was created with Spire.PDF for Python.
TELK OMNIKA T elecommun Comput El Control ❒ 1725 to end. The number of nodes is v aried from 50, 100, 150, 200, and 250 with the delay . Here where the node increases, the ti me tak en to transmit the pack et also increases. So, Figure 4 sho ws the proposed rey CGSTEB with the SEA G g ate w ay will be increased in time to deli v er the pack et from end-to-end nodes. The higher end- to-end delay in the proposed p r otocol is lik ely due to a combination of node increase, routing comple xity , and possibly ener gy-ef cient strate gies. Despite this, the protocol is well-suited for applications where ener gy ef cienc y , reliability , or non-time-critical data transmission is more important than immediate data deli v ery . Finally , the delay of the planned mo v e to w ards 90% higher than CGSTEB and rey CGSTEB. Observing Figure 5, e ner g y consumption describes the entire quantity of ener gy inspired by the nodes to broadcast the frame. Here the ener gy v aries in rounds. It is measured by Joules. Figure 5 represents the graphical representation of ener gy . When the round increases, the ener gy decreases. Here, the proposed rey CGSTEB sho ws ener gy consumption. Here the graph X-axis represents the number of rounds and Y -axis represent ene r gy measured by Joules. The illustration of ener gy for approach wit h dif ferent round consequences. It is accomplished that the ener gy consumption for proposed rey CGSTEB with a SEA G approach of 89% is lesser than the e xisting. In the fourth e xperiment, we implement to e v aluate the netw ork throughput. The number of sensor nodes is v aried from 50, 100, 150, 200, and 250. The simulation result is sho wn in Figure 6. In where, the number of nodes increases, the throughput will increase. Although throughput and delay ha v e an in v erse relationship, under some circumstances, both might increase at the same time due to a v ariety of causes including congestion, retransmissions, and netw ork conditions. Ho we v er , the throughput of our proposed approach is higher 76% than e xisting one. Figure 5. Comparison of ener gy consumption Figure 6. Comparison of netw ork throughput Escalating QoS by r ey optimization of CGSTEB r outing pr otocol with ...(R. Madonna Arieth) Evaluation Warning : The document was created with Spire.PDF for Python.
1726 ❒ ISSN: 1693-6930 W e summarizes the performance of the proposed Firey-CGSTEB with SEA G prot ocol compared with CGSTEB, Firey-CGSTEB, PSO-GSTEB, and LEA CH. The results clearly demonstrate the adv antages of our approach: pack et loss w as reduced by 86.2%, ener gy consumpti on decreased by 76.2%, netw ork lifetime increased by 88.7%, and throughput impro v ed by 72.4% relati v e to CGSTEB. These impro v ements are consis- tent across 30 independent simulation runs, with mean v alues and standard de viations reported. The substantial g ains highlight the ef fecti v enes s of combining Firey-based cluster head select ion with SEA G-assisted routing, particularly in e xtending netw ork lifetime and ensuring reliable data deli v ery . Ov erall, the results conrm that the proposed Firey-CGSTEB with SEA G frame w ork subs tantially impro v es QoS metrics in WSNs. The impro v ements in pack et deli v ery , ener gy ef cienc y , and lifetime are statistically signicant and rob ust, while the increase in end-to-end delay represents a manageable trade-of f de- pending on the application domain. This mak es the protocol particularly well-suited for smart city monitoring, healthcare sensing, and industrial automation, where long-term stability and reliability are more critica l than real-time response. 5. CONCLUSION The clusteri ng GSTEB i s non-deterministic polynomial-time hard (NP-hard) in nature. Here inte gra- tion of rey optimization with the SEA G enhances the o v erall performance of WSNs by ef fecti v e selecting the cluster head and formation of cluster for data trans mission. Based on tness v alue, rey optimization selects the most suitable cluster heads and there by it balances the intra-cluster ener gy consumption and pro- longing netw ork lifetime. The introduction of SEA G between the cluster heads and the base station reduces the communication head b urden, reduces pack et loss and delay , and ensures f aster data transmission, leading to impro v ed QoS compared to e xisting approaches such as clustered GSTEB and standalone rey clustering. The k e y contrib utions of this study lie in optimizing ener gy ef cienc y , e xtending netw ork usage, and achie ving more stable data routi ng. Despite this, there are some constraints, such as the computational o v erhead of the metaheuristic optimization process and practical dif culties in real-w orld scenario. In the future, the focus will be on v alidating the approach through real-w orld testbeds and e xamining h ybrid optimization methods to reduce comple xity and strengtheni n g SEA G nodes with lightweight security m echanisms to safe guard ag ainst potential vulnerabilities. FUNDING INFORMA TION Authors state no funding in v olv ed. A UTHOR CONTRIB UTIONS ST A TEMENT This journal uses the C ontrib utor Roles T axonomy (CRediT) to recognize indi vidual author contrib u- tions, reduce authorship disputes, and f acilitate collaboration. Name of A uthor C M So V a F o I R D O E V i Su P Fu R. Madonna Arieth ✓ ✓ ✓ ✓ ✓ ✓ ✓ Ramya Go vindaraj ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ Subrata Cho wdhury ✓ ✓ ✓ ✓ ✓ ✓ ✓ Thi Thu Nguyen ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ Duc-T an T ran ✓ ✓ ✓ ✓ ✓ C : C onceptualization I : I n v estig ation V i : V i sualization M : M ethodology R : R esources Su : Su pervision So : So ftw are D : D ata Curation P : P roject Administration V a : V a lidation O : Writing - O riginal Draft Fu : Fu nding Acquisition F o : F o rmal Analysis E : Writing - Re vie w & E diting CONFLICT OF INTEREST ST A TEMENT Authors state no conict of interest. TELK OMNIKA T elecommun Comput El Control, V ol. 23, No. 6, December 2025: 1718–1728 Evaluation Warning : The document was created with Spire.PDF for Python.
TELK OMNIKA T elecommun Comput El Control ❒ 1727 D A T A A V AILABILITY The data that support the ndings of this study are a v ailable from the corresponding author , [N.T .T], upon reasonable request. REFERENCES [1] Y . Liu, Q. W u, T . Zhao, Y . T ie, F . Bai, and M. Jin, “ An impro v ed ener gy-ef cient routing protocol for wireless sensor netw orks, ” Sensor s (Switzerland) , v ol. 19, no. 20, p. 4579, Oct. 2019, doi: 10.3390/s19204579. [2] C. Nakas, D. Kandris, and G. V isv ardis, “Ener gy Ef cient Routing in W ireless Sensor Netw orks: A Comprehensi v e Surv e y , ” Algorithms , v ol. 13, no. 3, p. 72, Mar . 2020, doi: 10.3390/a13030072. [3] R. K. V arun and R. C. Gangw ar , “ Hierarchical ener gy ef cient routing in wireless sensor netw orks and its challenges, ” International J ournal of Engineering and Advanced T ec hnolo gy , v ol. 9, no. 1, pp. 4024–4027, Oct. 2019, doi: 10.35940/ijeat.A1173.109119. [4] K. Selv akumar and R. M. Arieth, “Contrast for QOS based clustered ener gy ef cient protocol with PSO and multi-hop g ate w ays in wireless sensor netw ork, ” Cluster Computing , v ol. 22, no. S5, pp. 11883–11890, Sep. 2019, doi: 10.1007/s10586-017-1506-z. [5] G. Quoc-Anh, N. Dinh-Chinh, T . Duc-Nghia, T . Duc-T an, K. T . Nguyen, and K. Sandrase g aran, “W ireless technology for monitor - ing site-specic landslide in V ietnam, ” International J ournal of Electrical and Computer Engineering (IJECE) , v ol. 8, no. 6, pp. 4448–4455, 2018, doi: 10.11591/ijece.v8i6.pp4448-4455. [6] W . Ab ushiba, P . Johnson, S. Alharthi, and C. Wright, “ An ener gy ef cient and adapti v e clustering for wireless sensor netw ork (CH- leach) using leach protocol, ” in ICENCO 2017 - 13th International Computer Engineering Confer ence: Boundless Smart Societies , IEEE, Dec. 2017, pp. 50–54, doi: 10.1109/ICENCO.2017.8289762. [7] N. Gharaei, K. Ab u Bakar , S. Z. M. Hashim, and A. H. Pourasl, “Inter - and intra-cluster mo v ement of mobile sink algo- rithms for cluster -based netw orks to enhance the netw ork lifetime, ” Ad Hoc Networks , v ol. 85, pp. 60–70, Mar . 2019, doi: 10.1016/j.adhoc.2018.10.020. [8] A. Fujimoto, Y . Masui, T . Y oshihiro, and F . Uchio, “Beacon scheduling in re cei v er -initiated MA C protocols for lo w- delay and ener gy-ef cient WSNs, ” IEEE W ir eless Communications and Networking Confer ence , WCNC , 2017, doi: 10.1109/WCNC.2017.7925660. [9] P . R. Sujae and S. Arulselvi, “ An ef fecti v e intra and inter cluster formation with scheduling technique for a WSN system, ” Interna- tional J ournal of Inno vative T ec hnolo gy and Exploring Engineering , v ol. 8, no. 8, pp. 2399–2404, 2019. [10] S. Kaur and R. Mahajan, “Hybrid meta-heuristic optimization based ener gy ef cient protocol for wireless sensor netw orks, ” Egyp- tian Informatics J ournal , v ol. 19, no. 3, pp. 145–150, 2018, doi: 10.1016/j.eij.2018.01.002. [11] N. C. Minh et al. , “Cross-layer design for wireless sensor netw orks using cooja tool in contiki operating system, ” Lectur e Notes in Networks and Systems , v ol. 471 LNNS, pp. 484–494, 2022, doi: 10.1007/978-981-19-3394-3 55. [12] K. M. A w an, A. Ali, F . Aadil, and K. N. Qureshi, “Ener gy ef cient cl uster based routing algorithm for wireless sensor s net- w orks, ” 2018 Int ernational Confer ence on Advancements in Computational Sciences , ICA CS 2018, v ol. 2018, pp. 1–6, 2018, doi: 10.1109/ICA CS.2018.8333486. [13] H. Oudani et al. , “Ener gy Ef cient in wireless sensor netw orks using cluster -based approach routing, ” International J ournal of Sensor s and Sensor Networks. Special Issue: Smart Cities Using a W ir eless Sensor Networks , v ol. 5, no. 1, pp. 6–12, 2017, doi: 10.11648/j.ijssn.s.2017050501.12. [14] M. Dener , “ A ne w ener gy f f cient hierarchical routing protocol for wireless sensor netw orks, ” W ir eless P er sonal Communications , v ol. 101, no. 1, pp. 269–286, 2018, doi: 10.1007/s11277-018-5687-4. [15] J. P . Mishra, K. Singh, and H. Chaudhary , “Research adv ancements in ocean en vironmental monitoring systems using wireless sensor netw orks: a re vie w , ” T elk omnika (T elecommunication Computing Electr onics and Contr ol) , v ol. 21, no. 3, pp. 513–527, 2023, doi: 10.12928/TELK OMNIKA.v21i3.24010. [16] A. K. Ab u-Bak e, “Ener gy-ef cient routing in cluster -based wireless sensor netw orks: optimization and analysis, ” J or dan J ournal of Electrical Engineering , v ol. 2, no. 2, p. 146-159, 2016. [17] Z. Han, J. W u, J. Zhang, L. Li u, and K. T ian, “ A general self-or g anized tree-based ener gy-balance routing protocol for wireless sensor netw ork, ” IEEE T r ansactions on Nuclear Science , v ol. 61, no. 2, pp. 732–740, 2014, doi: 10.1109/TNS.2014.2309351. [18] S. Dehghani, M. Pourzaferani, and B. Barekatain, “Comparison on ener gy-ef cient cluster based routing algorithms in wireless sensor netw ork, ” Pr ocedia Computer Science , v ol. 72, pp. 535–542, 2015, doi: 10.1016/j.procs.2015.12.161. [19] K. Shah, T . Maity , Y . S. Dohare, D. T yagi, D. Rathore, and D. S. Y ada v , “ICIC: a dual mode intra-cluster and inter -cluster ener gy minimization approach for multihop WSN, ” IEEE Access , v ol. 10, pp. 70581–70594, 2022, doi: 10.1109/A CCESS.2022.3188684. [20] D. W ang, D. T an, and L. Liu, “P article sw arm optimization algorithm: an o v ervie w , ” Soft Computing , v ol. 22, no. 2, pp. 387–408, 2018, doi: 10.1007/s00500-016-2474-6. [21] A. Ahad, Z. Ullah, B. Amin, “Comparison of ener gy ef cient routing protocols in wireless sensor netw ork, ” American J ournal of Networks and Communications , v ol. 6, no. 4, p. 67, 2017, doi: 10.11648/j.ajnc.20170604.12. [22] P . K uila and P . K. Jana, “Ener gy ef cient cl ustering and routing algorithms for wireless sensor netw orks: particle sw arm optimiza tion approach, ” Engineering Applications of Articial Intellig ence , v ol. 33, pp. 127–140, Aug. 2014, doi: 10.1016/j.eng appai.2014.04.009. [23] S. R. Deepa and D. Rekha, “Cluster optimization in wireless sensor netw ork based on optimized Articial Bee Colon y algorithm, ” IET Networks , v ol. 10, no. 6, pp. 295–303, No v . 2021, doi: 10.1049/ntw2.12023. [24] P . C. S. Rao, P . K. Jana, and H. Banka, “ A particle sw arm opti mization based ener gy ef cient cluster head selection algorithm for wireless sensor netw orks, ” W ir eless Networks , v ol. 23, no. 7, pp. 2005–2020, Oct. 2017, doi: 10.1007/s11276-016-1270-7. [25] A. Mahboub, M. Arioua, and E. M. En-Naimi, “Ener gy-ef cient h ybrid K-means algorithm for clustered wireles s sensor net- w orks, ” International J ournal of Electrical and Computer Engineering (IJECE) , v ol. 7, no. 4, pp. 2054–2060, 2017, doi: 10.11591/ijece.v7i4.pp2054-2060. [26] K. Mehta and R. P al, “Ener gy ef cient routing protocols for wireless sensor netw orks: a surv e y , ” International J ournal of Computer Applications , v ol. 165, no. 3, pp. 41–46, 2017, doi: 10.5120/ijca2017913840. Escalating QoS by r ey optimization of CGSTEB r outing pr otocol with ...(R. Madonna Arieth) Evaluation Warning : The document was created with Spire.PDF for Python.