Performance Evaluation of Nsga-iii Based Energy Efficient Protocol for Wireless Sensor Networks

Energy efficiency has recently turned out to be primary issue in wireless sensor networks. Sensor networks are battery powered, therefore become dead after a certain period of time. Thus, improving the data dissipation in energy efficient way becomes more challenging problem in order to improve the lifetime for sensor devices. The clustering and tree based data aggregation for sensor networks can enhance the network lifetime of wireless sensor networks. Non-dominated Sorting Genetic Algorithm (NSGA) -III based energy efficient clustering and tree based routing protocol is proposed. Initially, clusters are formed on the basis of remaining energy, then, NSGA-III based data aggregation will come in action to improve the intercluster data aggregation further. Extensive analysis demonstrates that proposed protocol considerably enhances network lifetime over other techniques. Keywords: Wireless Sensor Networks, Ant Colony Optimization, Energy Efficient, Particle swarm optimization. Cite this Article: Er. Harjot Kaur, Dr. Gaurav Tejpal and Dr. Sonal Sharma, Performance Evaluation of NSGA-III Based Energy Efficient Protocol for Wireless Sensor Networks, International Journal of Mechanical Engineering and Technology 8(9), 2017, pp. 325–336. http://www.iaeme.com/IJMET/issues.asp?JType=IJMET&VType=8&IType=9 1. INTRODUCTION Modern progresses in digital electronics [2], micro-electro-mechanical system, and wireless communications have empowered the growth of small-sized sensor nodes, which have lowpower, low-cost and are multifunctional. These sensor nodes have capability to sense and communicate. Wireless sensor networks [1] are made up of a large number of sensor nodes, Er. Harjot Kaur, Dr. Gaurav Tejpal and Dr. Sonal Sharma http://www.iaeme.com/IJMET/index.asp 326 [email protected] densely deployed either inside the region or very near to it. Working of WSN is shown in Figure 1. Energy conservation is the major matter in wireless sensor network. Limited power nodes which cannot be replaced can be carried by sensor nodes. In WSN, sensor nodes sense data and transmit it to the base station. Since data from neighbouring sensor nodes [3] may be redundant, it becomes complex for base station to process large amount of data. Moreover, sensor nodes have their own energy. Due to redundant transmissions and loss of energy, lifetime of sensor nodes can decrease. To increase lifetime, data aggregation [4] is performed. Data aggregation means to collect and aggregate data [5] from multiple sensors to eliminate redundancy and conserve energy. Figure 1 Working of Wireless sensor network Wireless sensor networks have a wide range of applications in areas [2] such as security, military and health. For instance, a doctor can monitor the physiological data about a patient remotely. The current health condition of the patient is better understood by the doctor. Foreign chemical agents can be detected in the air and water with the help of sensor network. Pollutant’s type, amount and location can be identified. Lin et al. (2015) [4] utilized evolutionary game theory to select CH to reduce the hot-spot problem. In this method, size of cluster is optimized through optimal cluster size algorithm. The appropriate selection of CHs reduces the energy consumption and enhances the life of network. Gong et al. (2015) [5] designed a routing protocol ETARP (i.e., Energy Efficient Trust-Aware Routing Protocol for Wireless Sensor Networks) to reduce the energy consumption and increase the security during communication among nodes in WSNs. The selection of route between sensor nodes is based on utility theory. Shi et al. (2015) [6] addressed the issue of mobile sinks like route maintenance in WSNs by introducing dynamic layered routing protocol. The distribution frequencies and scopes of routing updates are minimized using the combination of dynamic anchor selection and dynamic layered Voronoi scoping. Performance Evaluation of NSGA-III Based Energy Efficient Protocol for Wireless Sensor Networks http://www.iaeme.com/IJMET/index.asp 327 [email protected] Leu et al. (2015) [7] utilized Regional Energy Aware Clustering with Isolated Nodes (REAC-IN) algorithm to select CHs based on weight. Weight is calculated considering each sensor’s residual energy and regional average energy of every sensor in all clusters. Shen et al. (2015) [8] solved the problem of delay in message transmission in underwater WSNs using Location-Aware Routing Protocol (LARP). In this method, position knowledge of sensor nodes is used to facilitate message transmission. Bouyer et al. (2015) [9] used fuzzy C-means (FCM) algorithm to create optimum number of CHs in LEACH algorithm to reduce the energy and prolong the network life-time. Cai et al. (2015) [10] proposed Bee-Sensor-C routing protocol inspired from Bee Sensor (i.e. bee-inspired routing protocol) that can form clusters dynamically and transmit the data in parallel fashion. Shankar et al. (2016) [11] used hybrid Particle Swarm Optimization (PSO) and Harmony Search Algorithm (HSA) to select CH efficiently utilizing minimum energy. Zahedi et al. (2016) [12] presented the problem of uneven distribution of CHs, unbalanced clustering, and their scope to limited applications of WSNs. They used fuzzy c-means clustering algorithm to create balanced clusters and Mamdani fuzzy inference system to select suitable CHs. Fuzzy rules are optimized through swarm intelligence algorithm based on firefly algorithm. Sabet and Naji (2016) [13] implemented the multi-level route-aware clustering (MLRC) technique to save energy in decentralized clustering protocols. The main advchromosomeage of this protocol is that it creates a cluster and routing tree, simultaneously, to reduce an unnecessary generation of routing control packets. Naranjo et al. (2017) [14] presented ProlongStable Election Protocol (P-SEP) to elect the CHs among heterogeneous nodes in fog-supported WSNs to increase the life of network. Xenakis et al. (2017) [15] utilized simulated annealing technique to control the topology by maximizing the network coverage and lifetime of WSNs as objective functions. Nayak and Vathasavai (2017) [16] utilized type-2 fuzzy logic in WSNs to make a decision for CH efficiency. Ouchitachen et al. (2017) [17] implemented IMOWCA (Improved Multi-Objective Weighted Clustering Algorithm) for the selection of CHs. Residual energy is used to select the best performing node for further communication with BS. Base Station Genetic Algorithm is utilized to balance the energy among different clusters. Elshrkawey et al. (2017) [18] addressed the issues of LEACH protocol like improper selection of CH, formation of unbalanced clusters, and continuous transmission of updating data. They used threshold value to elect CHs, sensor nodes send their updated data in their allotted time, and modified TDMA scheduling is utilized to break steady state phase. Rani et al. (2017) [19] used E-CBCCP protocol to cache the data at CH and relay node to evade the communication of same data packets. Control packets are used to inform all sensor nodes that data packets are same and do not transmit the data packets. Laouid et al. (2017) [20] designed an approach to select the best route based on hop count and residual energy of each sensor node to maximize the life of network. Ez-zazi et al. (2017) [21] utilized adaptive coding scheme considering channel state and distance between inter nodes to scrutinize the trade-off between energy efficiency and reliability. Huang et al. (2017) [22] used public transportation vehicles as mobile sinks to gather data. To balance the energy consumption, an energy-aware routing and energy-aware unequal clustering algorithms are used. Zhao et al. (2017) [23] utilized layer-based diffusion particle swarm optimization approach to optimize the position of sink and sensor to sink route to maximize the lifetime of WSNs. In this paper, we propose improved method for General Self-Organized Tree based Energy Balance routing protocol (tree-based). In present tree-based protocol routing tree is manufactured where tree centered routing is performed to transmit knowledge to the bottom section however in that if the parent node dies the topography must be repair again that'll Er. Harjot Kaur, Dr. Gaurav Tejpal and Dr. Sonal Sharma http://www.iaeme.com/IJMET/index.asp 328 [email protected] consume a lot of power and there might be loss of knowledge also. To prevail around the problem of sign delay and knowledge reduction in the system because of the nodes disappointment in the root to sink, cluster based aggregation process can be utilized. In big system, wellorganized sign of knowledge to the sink requires obtaining the maximum route according to how many trips; therefore, knowledge can be aggregated at group head which is to be transmitted to the bottom station. The clustering strategy may minimize knowledge redundancy and reduce the congestive routing traffic in knowledge transmission. Following the clustering tree centered routing at the cluster-heads it is required to obtain the shortest route between the source and the sink, but the smallest route issue is NP-Hard in nature [22]. 1.1. Contribution: Following are our main contributions in this research paper: 1. First of all, we have evaluated the performance of some well-known existing energy efficient protocols for WSNs. 2. Based upon the comparative analysis we have found that effective inter-cluster data aggregation using met heuristic techniques can improve the network lifetime further. 3. We have designed and implemented well-known NSGA-III based clustering treebased protocols to enhance the results further. 4. Extensive analysis has also been done to evaluate the effectiveness of the proposed technique. Rest of the paper is organized as follows: In Section 2, network energy model is described for WSNs. Section 3, describes the proposed technique with suitable mathematical formulation. Experimental Set-up and results are demonstrated in Section 4. Concluding remarks are demonstrated in Section 5. 2. NETWORK ENERGY MODEL In this research work, we have randomly deployed WSN with “N” sensor nodes in M*N network field. All nodes even including the sink are stationary in nature. Each node has its own unique identification number. Each node monitors the given environment and communicates data with sink. Whenever communication is done given node have to spent some energy based upon the distance (D) with sink. All the communication links are symmetric in nature. 2.1. Energy Model Whenever a node sends or receives it has to spend some energy based upon two channel propagation models called free space (D power loss) for the purpose of one-hopordirect transmission and the multipath fading channel (D power loss) for packet transmission via multihop. Therefore, energy consumption model can be mathematically defined as follows: E T L, D LE + Lε D , D < D , E + Lε D D ≥ D , (1) Here?? Is the size of data packet, ε is free space energy loss, ??mp is multipath energy loss. D , Is a threshold distance which determine which energy model will be used. It can be calculated as follows: D = ! "#$% & "' (2) Performance Evaluation of NSGA-III Based Energy Efficient Protocol for Wireless Sensor Networks http://www.iaeme.com/IJMET/index.asp 329 [email protected] 2.2 CH formation In this section level-based clustering will be discussed. CHs are formed using energy aware threshold function. Which means nodes who has more energy will have more probability to become CHs. Each node generates random value and try to become CH. If random value is less than evaluated Threshold (T(i)), then it will become CH, become member node otherwise. T(i) can be mathematically evaluated as follows ( ) = * +,./,+,-01.3+4 5 . ,+,67 ∗ 9) 1 9:;< 1 For all nodes if E = r >0 (3) Here r represents the current round in WSNs network lifetime, E = r is the current energy of given node i. E@AB Represents average remaining energy which is evaluated using eqn. (4) E@AB = ∑DE F for every node i (4) N is the number of total nodes. 3. NSGA-III BASED TREE-BASED TECHNIQUE In this section, we propose an NSGA-III-tree-based based routing to develop shortest path among available CHs and sink. NSGA-III is a well-known metaheuristic technique which can find optimal path between given set of nodes with sink as destination. The actual design in the proposed NSGA-III may be identified inside Algo 1. Initial, a collection of reference point is created, which is denoted G = HIJ, I ,...ILM.For an N objective problem, IO PQ 1,2. . . T is an N dimensional vector represented by IO, T = IOJ, IO ,....IOU VW , where IO,L ≥ X, T = 1,2. . , N and ∑ IO,L ULYJ = 1. Next, the original population using N users is definitely arbitrarily produced. Intended for an ideal point Q∗ it is somtimes so time − consuming in order to determine exact, Qh∗ , therefore it is really estimated with the bare minimum cost discovered to date to get intent objective ih̅, and is kept up to date in the search. Actions 5-21 are generally iterated before the broadcasting is satisfied. In Step 6, an offspring population klW is created by utilizing the same genetic operators with those in NSGA-III. Along with such as, klW it is integrated with e present population mlW as well as form a different population. nlW .Thereafter,nlW . Is normalized applying an ideal point Q∗. Following normalization, the particular clustering user is used to split the particular users inside nlW into a set of T clusters oOpJ, oOp where the cluster oOpP is definitely depicted by the reference point IL. Then, a non-dominated organizing dependant on importance (not Pareto-dominance) works to help classify nlW in unique q non-domination levels iJ, i , rNs tu uN . Dominance, which is a key principle in q NSGA-III, will be presented later. The moment non-dominated-sorting have been accomplished, right now their maiming steps complete the population slots inmvwlWxJ making use of 1 level at one time, commencing fromih. Compared with both equally NSGA-II and NSGA-III, we just at random select answers in the last recognised level ih in q NSGA-III, due to the fact dominance offers stressed out both equally unity plus diversity. Certainly, several tactics to enhance the diversity could also be used likewise in Step 18. Er. Harjot Kaur, Dr. Gaurav Tejpal and Dr. Sonal Sharma http://www.iaeme.com/IJMET/index.asp 330 [email protected] Algorithm 1 Proposed NSGA-III based path selection 1: G ← z{N{|r}{ ~{i{|{No{ muN}t 2: m GN}rP€{ mu‚Pr}uN 3: ƒ ← GN}rP€{ Gs{rP muN} 4: }́ ← q 5: while the termination criterion is not met do 6: klW Make Offspring Population (mvw }́) 7: …sr}{ Gs{rP muN} klW 8: nlW ← mvwlW ∪ klW : Normalize ( nl, W Q∗) 10: oOp ← Clustering ( nlW Q∗ 11: iJ, i ... } ← q Non-dominated-sort. (nlW , oOp) 12: mvw }́ + 1) ; 13: ˆ ← 1 14: while ǀmvw }́ + 1ǀ + ǀ iŠ ≤ Œ do 15: mvw }́ + 1 ← mvw }́ + 1 ∪ iŠ 16 ˆ ← ˆ + 1