EACO and FABC to multi‐path data transmission in wireless sensor networks
2016; Institution of Engineering and Technology; Volume: 11; Issue: 4 Linguagem: Inglês
10.1049/iet-com.2016.0859
ISSN1751-8636
AutoresRajeev Kumar, Dilip Kumar, Dinesh Kumar,
Tópico(s)Security in Wireless Sensor Networks
ResumoIET CommunicationsVolume 11, Issue 4 p. 522-530 Research ArticleFree Access EACO and FABC to multi-path data transmission in wireless sensor networks Rajeev Kumar, Corresponding Author Rajeev Kumar rajeevkumares@gmail.com Department of Computer Science and Engineering, I.K.G. Punjab Technical University, Punjab, IndiaSearch for more papers by this authorDilip Kumar, Dilip Kumar Department of Electronics and Communication Engineering, SLIET Longowal, Punjab, IndiaSearch for more papers by this authorDinesh Kumar, Dinesh Kumar Department of Information Technology, D.A.V. Institute of Engineering and Technology, Jalandhar, Punjab, IndiaSearch for more papers by this author Rajeev Kumar, Corresponding Author Rajeev Kumar rajeevkumares@gmail.com Department of Computer Science and Engineering, I.K.G. Punjab Technical University, Punjab, IndiaSearch for more papers by this authorDilip Kumar, Dilip Kumar Department of Electronics and Communication Engineering, SLIET Longowal, Punjab, IndiaSearch for more papers by this authorDinesh Kumar, Dinesh Kumar Department of Information Technology, D.A.V. Institute of Engineering and Technology, Jalandhar, Punjab, IndiaSearch for more papers by this author First published: 01 March 2017 https://doi.org/10.1049/iet-com.2016.0859Citations: 16AboutSectionsPDF ToolsRequest permissionExport citationAdd to favoritesTrack citation ShareShare Give accessShare full text accessShare full-text accessPlease review our Terms and Conditions of Use and check box below to share full-text version of article.I have read and accept the Wiley Online Library Terms and Conditions of UseShareable LinkUse the link below to share a full-text version of this article with your friends and colleagues. Learn more.Copy URL Share a linkShare onFacebookTwitterLinkedInRedditWechat Abstract Owing to the extensive growth of wireless technology for sending and collecting a variety of information for the different applications, routing is a major challenge to find the optimal path for the data transmission. In this study, the authors have developed a new algorithm called, exponential ant colony optimisation (EACO) to route discovery problem in wireless sensor network after finding the cluster heads (CHs) using fractional artificial bee colony (FABC) algorithm. In the first step, CHs are found out using the FABC algorithm with fitness function considering the distance, energy and delay. In the second phase, ACO algorithm is modified with exponential smoothing model for multi-path route discovery. This new algorithm called, EACO found the optimal routes among CHs to transmit a data from any source node to base station with multiple objectives including energy, distance, intra-cluster delay and intercluster delay. These objectives are effectively formulated as new fitness function to find the optimal route path. From the experimentation, the outcome showed that the cumulative energy kept after 2000 round of experiments is 0.2039 for the proposed algorithm but the existing approach (threshold + ACO) kept only 0.0338. 1 Introduction Wireless sensor networks (WSNs) [1] find a lot of interest in various fields where collecting important data in remote or inaccessible locations is required. In WSNs, designing of routing protocols is indispensable one to reliant and design goals for different applications [2, 3]. When analysing the literature to find the availability of routing protocol for reducing energy, optimisation-based clustering protocols are found to be more popular techniques to perform routing of data packets with high-energy efficiency. Recent research has proven that both multi-path and clustering communication are very efficient routing methods in WSNs. Clustering has been widely used to extend the network lifetime and achieve network scalability. Data from different sources in cluster are aggregated by reducing redundancy with the purpose of minimising energy consumption in transmission. On the other hand, in multi-path mechanism, two or more paths are established from source to destination. Multi-path routing can distribute the forwarding of the data packets across multiple paths, so that all nodes can utilise their batteries at a comparable rate, which contributes to prolonging the network lifetime and achieving load balance [4]. In clustering-based multi-path routing, every sensor node must belong to one and only one cluster. Sensor nodes send their sensed data to their related cluster heads (CHs). CHs next cumulate them and send it to an isolated base station through different CHs in multi-hop communication [5]. Multi-path routing can cut down the need for line updates, balance the traffic load and increase the data transfer rate in a wireless sensor network, improving the usage of the reserved energy of sensor nodes. The aim behind traffic spreading is that for a given total energy consumption in the network, at each moment, every node should have spent the same amount of energy. The objective is to assign more loads to under-utilised paths and less load to over-committed paths, so that uniform resource utilisation of all available paths can be ensured. Multi-path routing is cost effective for heavy load scenario, while a single path routing scheme with a lower complexity may otherwise be more desirable [6]. In this regard, designing efficient routing protocols to establish high-quality multi-hop paths from each event area toward the sink is one of the most important issues in developing wireless sensor network [7]. This paper aims at achieving multi-path routing in wireless sensor network using the fractional artificial bee colony (FABC) algorithm [8] and the proposed exponential ant colony optimisation (EACO) algorithm. At first, FABC algorithm is used to find the CHs after minimising the objectives include distance, energy and delay. Once the CHs are found out, EACO algorithm is used to find the route path to send the data through CH to base station. Here, a new optimisation algorithm called EACO is developed. This new algorithm uses an exponential weighted moving average model to control the convergence rate of the ACO algorithm. This new algorithm utilises a new fitness function to evaluate the route discovery. Here, four objectives including energy consumption, distance travelled, intra-cluster and inter-cluster delay are considered to minimise the overall objective. The main contributions given in this work are as follows: A new optimisation algorithm called EACO is developed in this paper by extending the ACO algorithm given in [9]. This new algorithm uses an exponential weighted moving average model to control the convergence rate of the ACO. A new fitness function is proposed to evaluate the optimal route discover problem. Here, three objectives include energy consumption, distance travelled, intra-cluster and inter-cluster delay are considered to minimise the overall objective. This paper is organised as follows. Section 2 presents review of related work and Section 3 discusses system model and problem statement. Section 4 presents the proposed EACO and FABC to multi-path data transmission in WSNs. Section 5 presents the experimental results and evaluation. Finally, conclusion is given in Section 6. 2 Review of related works Literatures [10–15] present various energy aware routing protocols for WSN. One of the mostly used algorithms for energy aware routing is low-energy adaptive clustering hierarchy [16]. Then, particle swarm optimisation (PSO) algorithm which is benchmark optimisation algorithms is applied for routing in WSN [17] and evolutionary-based routing protocol can be found in [18]. Very recently, harmony search algorithm [19], which is one of the optimisation approaches used as metaheuristic optimisation method, is used as clustering-based routing protocol for energy-efficient wireless sensor network. ABC algorithm is recently applied for CH selection for energy reduction in WSN [20] and subsequently hybrid ABC algorithm is also applied for routing to extend WSN lifetime [21]. Degree-energy-based local random routing strategy is proposed in [22] which considers degree-energy-awareness principle to perform good on costs of various sensor network. Han et al. [23] presented self-organised tree-based energy-balance routing protocol to minimise total energy consumption but also to balance WSN load. Amgoth and Jana [5] presented an algorithm based on a clever strategy of CH selection using residual energy. It efficiently forms the directed virtual backbone of CHs to facilitate data routing to the sink. When compared with these CH selection algorithms, this paper considers FABC algorithm which contains multiple criteria to select the optimal CHs. Nowadays, multi-path routing approach is introduced as an effective technique for improving sensor networks performance in terms of energy consumption, fault tolerance, reliability and throughput [24]. In [25], Ganesan et al. presented a disjoint multi-path routing based on local information, which is a distributed algorithm and can achieve load balancing. This algorithm uses a primary route to transmit data. Only when the primary route fails, the alternative route can be used. However, this algorithm is not attractive for the network lifetime. In [26], a meshed multi-path routing with efficient strategy has been described. Such an algorithm can achieve a better throughput than the traditional multi-path algorithms. Sharma and Jena [27] proposed a cluster-based multi-path routing protocol, which uses the clustering and multi-path techniques to reduce energy consumption and increase the reliability. Jin et al. [28] developed a practical passive cluster-based node-disjoint many to one multi-path routing protocol to satisfy the requirements of energy efficiency and quality of service (QoS) in practical WSNs. Zhong et al. [29] proposed a secure clustering and reliable disjoint multi-path route discovery method to transfer the data aggregation results. When compared with these multi-path routing algorithms, this paper presents the optimal criterion to select the CHs and path for routing mechanism. 3 System model and problem statement Network model: A wireless sensor network contains n sensor nodes along with one base station or sink node, . Every sensor node can be CHs, or normal node which are distributed within the data space of and in metres. The location of sink node is denoted as . Radio model: The radio model of the proposed system follows the model described in the previous paper [8] and the system follows a free space and multi-path fading model given in [30]. Problem statement: In wireless sensor network, the path of transmitting data from one sensor node to other sensor node should have less number of nodes because the data receiving and sending by every node would require energy. This can be controlled by selecting CHs from sensor node and CHs send the data to base station after collecting from their corresponding members. However, double path routing mechanism is not much effective if the nodes are distributed in big interval in the perspective of energy. So, the finding of multi-path to transmit the data has taken major challenge in this work. On the basis of these objectives, two major challenges are identified, the first one is to identify the CHs and the second challenge is to identify the paths. In these challenges, the formulation of objectives to select the CHs and the path should consider all the parameters involved. When a node is selected as CH, it should have more energy, less distance from their cluster member and less delay. When we formulate the path for data transmission, the path should have less distance from the sender to receiver, less energy consumption, less inter-cluster and intra-cluster delay. In [9], ACO algorithm was utilised to search for multiple paths after cluster formation. However, this algorithm [9], does not consider inter- and intra-cluster delays to formulate the paths. Also, the traditional ACO algorithm [9] does not consider the prior information of pheromone to find the paths. 4 Proposed EACO and FABC to multi-path data transmission in wireless sensor networks This section presents the proposed EACO and FABC to multi-path data transmission in WSNs. The proposed data transmission follows two stages of mechanisms. In the stages, sensor nodes are formed as groups and the CHs are selected from the groups using FABC algorithm. In the second stage, the path of data transmission is found out using the proposed EACO algorithm. The finding of data path is purely based on the four objectives such as, distance, energy, intra-cluster and inter-cluster delay. On the basis of these four objectives, the objective function is newly proposed and it is injected into the proposed EACO algorithm which is modified from the ACO algorithm including the exponential smoothing theory in pheromone updating process. Fig. 1 shows the block diagram of the proposed multi-path routing scheme. Figure 1Open in figure viewerPowerPoint Block diagram of the proposed method of multi-path routing 4.1 CH selection using FABC algorithm This step is to find out the CHs from a set of sensor nodes available in wireless sensor network. The CH selection mechanism can be formulated as searching problem as the aim is to identify the suitable sensor node to be activated as CHs. Accordingly, well-known search algorithm called ABC is integrated with the fractional theory to develop an algorithm called FABC [8]. The FABC algorithm given in the previous work [8] is then applied to wireless sensor network to optimally select the CHs. 4.2 Multi-path routing for data transmission using EACO algorithm This stage is to generate the data path for transmitting data from one sensor node to other sensor node. From the previous step, we identify the CH for all the sensor nodes. CHs are the first path node for all the sensor nodes. The second path to reach the base station is identified from the nodes which are selected as CHs. The subsequent paths are also selected from the CHs. So, sensor nodes transmit the data to its CHs and the cluster transmits the data to the base station through a set of CHs using four different constraints such as distance, energy, intra-cluster and inter-cluster delays. The path identification is done through the proposed EACO algorithm which is hybridisation of ACO algorithm with exponential weighted moving average [31]. The evaluation of path is done using the proposed objective function. 4.2.1 Proposed EACO algorithm The CHs selected from the previous step is directly passed through ACO algorithm [32] to obtain the multi-path construction. The ant system finds out the multi-paths for every CHs identified from previous steps. At first, the multi-paths for the first CH are identified through the proposed EACO algorithm. Simultaneously, the paths to reach base stations are identified for all other CHs by running the EACO algorithm again. Here, the ant system is built up with the size of , where is the number of ant locations which can be defined here as number of CHs selected from the previous stage. A is number of ants required to construct the solution or multi-path. Initially, ants are placed in a random way among the CHs and solution is constructed. For this solution, the proposed ant cost function is evaluated to identify the fitness of the solution. In the next iteration, the location of A artificial ants location should be updated to improve the solution. To update the solution, the pheromone update is necessary to increase the pheromone values associated with good solutions. In the traditional ACO algorithm, the pheromone associated with is updated as follows (1)where is the evaporation rate, A is the number of ants and is the quantity of pheromone laid on edge by ant k where Q is a constant and is the length of the solution constructed by ant k. In the proposed EACO algorithm, the updating of pheromone is modified using the exponential weighted moving average model. exponential weighted moving average (EWMA) [31] is one of the popular data prediction model. In this work, we integrated the EWMA model with the ACO model to further improve the searching behaviour. According to EWMA model, the prediction of pheromone can be formulated as (2)where e is the constant, is the output of the pheromone for the next iteration using EWMA model after and is the output of the pheromone of the last iteration using EWMA model. The above equation can be rewritten as (3)If we assume that the EWMA model predicts the output exactly as the original output, then we can replace the above equation to the original pheromone equation (4)The above formulae derived are used in the EACO algorithm for pheromone updating on every iteration. In the construction of a solution, ants select the following element to be visited through a stochastic mechanism. When ant k is in location i and has so far constructed the partial solutions p, the probability of going to place j is given by (5)where is the set of feasible components; that is, edges where l is a location not yet visited by the ant k. The parameters and control the relative importance of the pheromone versus the heuristic information , which is given by (6)where is the distance between two locations i and j. On the basis of the above computation, every ants location are updated for every iteration which will give a new path. This path is again evaluated based on the proposed cost function and the values are memorised. The same procedure is repeated for t number of iterations and the path which has the minimum cost value is selected as the optimal path for delivering the packet from source node to sink node. This process of finding path is repeated for all the CHs and the data communication can happen only through this path. Fig. 2 shows the algorithmic procedure of EACO algorithm. Figure 2Open in figure viewerPowerPoint EACO algorithm 4.2.2 Proposed cost function The evaluation of ant location is done using the proposed cost function, which considers the four objectives such as distance, energy, intra-cluster and inter-cluster delays. These four objectives are important in selecting the path of data transmission. These objectives are included the proposed cost function as follows (7)where , , and are weighted constants for the four objectives considered in the cost function. The cost function related to distance is represented as which considers the distance to be travelled as main parameter. The numerator of this cost function found out by taking the summation of the distance between two consequent paths. This distance value should be a minimum but the value should range in between 0 and 1. So, the denominator is found by doing the summation of all the maximum distance covered in between CHs. Now, the factor have a range in between 0 to 1 and 0 signifies the best case and 1 signifies the worst case (8) (9) (10)where is the distance between two consequent CHs and . The second parameter considered is energy which should be a maximum for the entire CHs selected in the path. Here, energy of every CH is divided with the maximum energy to obtain the value in between 0 and 1 and taking subtraction from maximum energy to convert the objective of minimisation. Now, the value of range between 0 to 1 and 0 signifies the best case and 1 signifies the worst case (11)where is the maximum energy of CH and q is the number of paths in the route R. The third parameter utilised here is intra-cluster delay. The intra-cluster delay mainly depends on the number of cluster member in the particular head. If more numbers of members under the head, then the delay is more. The objective here is to minimise the cluster member to make the system without any traffic. Accordingly, is divided by one and the value is then subtracted from 1. It signifies that the value 1 shows the worst case and 0 signifies the best case (12)The fourth parameter considered here is inter-cluster delay which heavily depends on the number of nodes to be traversed by the data packets. The value of is found out by taking the ratio of number of paths to be traversed and total number of nodes. Here, the best case provides the minimum value and the worst case signifies the maximum value (13) 4.2.3 Overall description Fig. 3 shows the sample example of description about the framework in multi-path route discovery. From this figure, we can easily understand that network contains 16 sensor nodes with one sink node. Every sensor node presented in the network wants to send a data to sink node. To accomplish this task, the first step is to find out the CH which is performed using FABC algorithm. This algorithm finds the nodes such as, 1, 3, 10 and 16 as CHs. The cluster member of CH 1 is 2, 5 and 6 and nodes 4, 7 and 8 under CH 3. Similarly, nodes 9, 13 and 14 are grouped under CH 10 and CH 16 has the nodes 11, 12 and 15 as cluster members. Then, every CHs has to obtain the path of traversal to reach base station. The route path of every CHs is identified by applying EACO algorithm. For example, CH 1 should send the data to base station through CH 3. This figure shows the complete route table for all the sensor nodes. From Fig. 3, sensor node 11 should send the data to the base station through CHs 16 and 10. Figure 3Open in figure viewerPowerPoint Overall description of the proposed method in route discovery 5 Results and discussion 5.1 Network simulation setup The simulation of wireless sensor network and the implementation of the proposed FABC algorithm are done using MATLAB. In the simulation, sensor nodes are fixed in the area of 100 m × 100 m and the base station is located in the middle of the region. Table 1 shows the parameters fixed for the experimentation of the proposed routing protocol. Table 1. Parameters fixed for the experimentation Parameters Variables Fixed values network parameters 100 100 n 50/100 b 4000 bit 0.5 10 pJ/bit/m2 0.0013 pJ/bit/m2 50 nJ/bit/m2 5 nJ/bit/signal FABC parameters 1 10 L 100 5 0.5 0.3 0.2 EACO parameters 0.5 A 10 Q 2.7 e 0.5 2 6 Fig. 4 shows the sample plot of the network for four different rounds of experiments. The simulation of the network environment is executed and the nature of sensor nodes for every round of experimented is visualised in Fig. 4. Fig. 4a shows the initial location of sensor nodes and base station. The black colour indicates the normal sensor nodes and star symbol indicates the CH. The cross symbol indicates the sink node. The red colour indicates the dead nodes. Figure 4Open in figure viewerPowerPoint Network simulation a r = 0 b r = 500 c r = 1000 d r = 2000 5.2 Parametric analysis of FABC + EACO algorithm The performance analysis of the proposed FABC + EACO algorithm is done by various parameters of , , and which are weighted constants used in the objective function. Here, three types of performances in terms of alive nodes, energy and goodput are analysed to find the optimal parameters suitable for the proposed method. Alive nodes are the total number of nodes alive over a period of time (round of experiments). Normalised network energy is the average energy retained in the network over a period of time (round of experiments). Goodput is the total number of bits delivered by the network to a destination per unit of time. (a) Performance analysis of life time of nodes: For performance analysis of life time of nodes, four various parametric values are fixed for , , and and life time of nodes are computed. The Fig. 5a shows the performance graph for 50 nodes and Fig. 5b shows the performance of 100 nodes. From the graph, we understand that when the value of , , and is fixed to 0.3, 0.5, 0.1 and 0.1, the performance shows the better as compared with other values for n = 100, and also the similar performance is achieved for 50 nodes for all the variety of weighted constants. (b) Performance analysis of energy consumption: The normalised energy for the network is computed after every round of experiments for various values of , , and . From Fig. 6a, we know that the better performance is achieved when the value of those constants are fixed to 0.5, 0.3, 0.1 and 0.1. The performance is more similar until the round of experiments is 600. After that, a small difference is there for all the combination of parametric values. Fig. 6b shows the normalised energy for the total of 50 nodes. In this graph, the performance is superior for the same values of weighted constants such as Fig. 6a. (c) Performance analysis of goodput: The important parameter taken for the analysis is goodput which is essential to prove the performance in routing mechanism. Here, four different kinds of values are fixed up for the weighted constant and the performance is monitored for 100 and 50 nodes. Figs. 7a and b show the performance graph for 100 and 50 nodes. From this figure, we understand that the similar performance is achieved until the round of experiments is around 1200 and the deviation can be found out after these rounds. However, the performance is good for the values of , , and to be fixed as 0.5, 0.3, 0.1 and 0.1. Figure 5Open in figure viewerPowerPoint Analysis of FABC + EACO algorithm in terms of alive nodes a n = 100 b n = 50 Figure 6Open in figure viewerPowerPoint Analysis of FABC + EACO algorithm in terms of normalised network energy a n = 100 b n = 50 On the basis of the performance analysis using life time, energy and goodput, the suggestion is that the proposed FABC + EACO performed better when the values of , , and is fixed as 0.5, 0.3, 0.1 and 0.1. 5.3 Comparative evaluation of the proposed FABC + EACO algorithm The comparative evaluation of the proposed FABC + EACO algorithm is performed with the ABC + ACO algorithm [20], threshold + ACO [9], GABC + ACO algorithm and PSO + ACO algorithm. Here, ABC + ACO algorithm is simulated such as the work in [20] for CH selection, and then routing was discovered such as the work given in [9]. Here, three parameters such as life time, energy and goodput are taken as objectives to evaluate the performance of the algorithms. (a) Life time: Fig. 8a shows the comparative analysis of FABC + EACO algorithm in terms of alive nodes for 100 nodes. From this figure, the performance is good for the ABC + ACO algorithm until rounds are equivalent to 1200. However, for the later rounds, the performance is much better for the proposed algorithm than the previous two algorithms. Fig. 8b shows the life time analysis of the proposed methods with the other two previous algorithms. Here, the proposed algorithm is much better as compared with other two algorithms. The proposed algorithm provides the top performance in 75% of rounds. (b) Energy consumption: Fig. 9a shows the comparative analysis of the proposed algorithm in terms of energy for 100 nodes. Here, we have obtained the top performance in 55% of the rounds. The threshold + ACO provides the worst performance than the ABC + ACO algorithm in most of the rounds. Fig. 9b visualises the comparison of the algorithms for 50 nodes. Here, the proposed algorithm is completely dominated and achieved the top performance in entire round of experiments. ABC + ACO is better than the threshold + ACO in most of the round of experiments. (c) Goodput: Fig. 9c shows the comparative analysis of three algorithms using goodput. From Fig. 9c, the proposed FABC + EACO obtained the top performance in all the round of experiments. Until the rounds of 1500, the proposed algorithms obtained the value of 1 and then the performance is degraded due to the dead nodes. The threshold + ACO performed better than the ABC + ACO in terms of goodput. Fig. 9d shows the performance of goodput for three algorithms when a number of sensor nodes are fixed as 50. Here, the performance of the proposed algorithm is much better than the existing two algorithms in all the round of experiments. Figure 7Open in figure viewerPowerPoint Analysis of FABC + EACO algorithm in terms on goodput a n = 100 b n = 50 Figure 8Open in figure viewerPowerPoint Comparative analysis of FABC + EACO algorithm in terms of alive nodes a n = 100 b n = 50 Figure 9Open in figure viewerPowerPoint Comparative analysis of FABC + EACO algorithm in terms of normalised network energy and goodput a Normalised network energy for n = 100 b Normalised network energy for n = 50 c Normalised goodput for n = 100 d Normalised goodput for n = 50 5.4 Discussion The descriptive performance of the five algorithms is given in Table 2. This table consists of the performance measure after experimenting with 2000 rounds. For 50 nodes, the proposed FABC + EACO outperformed in terms of three metrics such as alive node, energy and goodput. Here, the proposed algorithm keeps 44 nodes alive as compared with other algorithm. Similarly, the energy kept after 2000 round of experiments is 0.3458 for the proposed algorithm but the existing threshold + ACO kept only 0.0575. For 100 node analysis, the proposed FABC + ACO algorithm maintains the highest energy of 0.2039 which is high as compared with other algorithms. Also, the proposed algorithm maintains 73 nodes alive after 2000 round of experiments. Table 2. Descriptive comparison and improvement after 2000 rounds ABC + ACO FABC + EACO Threshold + ACO GABC + ACO PSO + ACO n = 50 alive node 29 44 8 36 18 energy 0.086 0.3458 0.0369 0.2159 0.0614 goodput 0.0563 0.0583 0.0575 0.0573 0.0569 n = 100 alive node 56 73 18 64 37 energy 0.1105 0.2039 0.0338 0.1572 0.0721 goodput 0.0293 0.0194 0.0292 0.0243 0.0292 The bold value indicates the better performance Through the better performance in terms of energy and goodput, the proposed EACO-based multi-path routing protocol is a special kind of routing protocol that mainly considers the energy management. The EACO-based multi-path routing protocol provides its own energy management using the optimal CH allocation. It performs the functions such as data transfer and reception, energy detection of current channel and link quality indication. Owing to these advantages of functions, the proposed algorithm can be easily adaptable to the IEEE 802.15.4 [33, 34] which is a low-rate communication protocol suitable for the systems having hardware in low complexity. 6 Conclusion We have presented the proposed EACO and FABC to multi-path data transmission in WSNs. At first, wireless senor network is simulated with defined network and radio model. Then, CHs are selected using the FABC algorithm considering three different objective constraints. Then, the CHs selected from the previous step is directly passed through the proposed EACO algorithm to obtain the multi-path construction. This new algorithm is hybridisation of exponential model and ACO algorithm. This ant system finds out the multi-paths for every CHs using the proposed cost function, which considers the four objectives such as distance, energy, intra-cluster and inter-cluster delays. This optimal route discovery strategy is evaluated using the energy, life time of nodes and goodput. The experimental results showed that the cumulative energy kept after 2000 round of experiments is 0.2039 for the proposed algorithm but the existing approach (threshold + ACO) kept only 0.0338. This work can be further enhanced by considering variety of QoS attributes to select the CHs and route discover problem. References 1Gautam N., Pyun J.-Y.: 'Distance aware intelligent clustering protocol for wireless sensor networks', J. Commun. Netw., 2010, 12, (2), pp. 122– 129 (doi: https://doi.org/10.1109/JCN.2010.6391368) 2Lee J.S., Cheng W.L.: 'Fuzzy-logic-based clustering approach for wireless sensor networks using energy predication', IEEE Sens. J., 2012, 12, (9), pp. 2891– 2897 (doi: https://doi.org/10.1109/JSEN.2012.2204737) 3Hammoudeh M., Newman R.: 'Adaptive routing in wireless sensor networks: QoS optimisation for enhanced application performance', Inf. Fusion, 2015, 22, pp. 3– 15 (doi: https://doi.org/10.1016/j.inffus.2013.02.005) 4Cai X., Duan Y., He Y. et al.: 'Bee-sensor-C: an energy-efficient and scalable multipath routing protocol for wireless sensor networks', Int. J. Distrib. Sens. Netw., 2015, 2015, p. 14, Article ID 976127 5Amgoth T., Jana P.K.: 'Energy-aware routing algorithm for wireless sensor networks', Comput. Electr. Eng., 2014, 41, pp. 357– 367 (doi: https://doi.org/10.1016/j.compeleceng.2014.07.010) 6Mishra N., Kumar A.: 'Comparative analysis: energy efficient multipath routing in wireless sensor network', Int. J. Comput. Sci. Mob. Comput., 2014, 3, (9), pp. 627– 632 7Radi M., Dezfouli B., Bakar K.A. et al.: 'IM2PR: interference-minimized multipath routing protocol for wireless sensor networks', Wirel. Netw., 2014, 20, (7), pp. 1807– 1823 (doi: https://doi.org/10.1007/s11276-014-0710-5) 8Kumar R., Kumar D.: 'Multi-objective fractional artificial bee colony algorithm to energy aware routing protocol in wireless sensor network', Wirel. Netw. (WINE), 2016, 22, (5), pp. 1461– 1474 (doi: https://doi.org/10.1007/s11276-015-1039-4) 9Yang J., Xu M., Zhao W. et al.: 'A multipath routing protocol based on clustering and ant colony optimization for wireless sensor networks', Sensors (Basel), 2010, 10, (5), pp. 4521– 4540 (doi: https://doi.org/10.3390/s100504521) 10Younis O., Fahmy S.: 'HEED: a hybrid, energy-efficient, distributed clustering approach for ad hoc sensor networks', IEEE Trans. Mob. Comput., 2004, 3, (4), pp. 660– 669 (doi: https://doi.org/10.1109/TMC.2004.41) 11Zhang B., Simon R., Aydin H.: 'Harvesting-aware energy management for time-critical wireless sensor networks with joint voltage and modulation scaling', IEEE Trans. Ind. Inf., 2013, 9, (1), pp. 514– 526 (doi: https://doi.org/10.1109/TII.2011.2174247) 12Yu M., Kin K.L., Ankit M.: 'A dynamic clustering and energy efficient routing techniques for sensor networks', IEEE Trans. Wirel. Commun., 2007, 6, pp. 3069– 3079 (doi: https://doi.org/10.1109/TWC.2007.06003) 13Fariborzi H., Luo M.R.C., Chen O.: 'Mobile sensor node deployment and asynchronous power management for wireless sensor networks', IEEE Trans. Ind. Electron., 2012, 59, (5), pp. 2377– 2385 (doi: https://doi.org/10.1109/TIE.2011.2167889) 14Ren F., Zhang J., He T. et al.: 'EBRA: energy-balanced routing protocol for data gathering in wireless sensor networks', IEEE Trans. Parallel Distrib. Syst., 2011, 22, pp. 2018– 2125 15Jiguo Y., Yingying Q., Guangui W. et al.: 'A cluster-based routing protocol for wireless sensor with non-uniform node distribution', Int. J. Electron. Commun., 2012, 66, pp. 54– 61 (doi: https://doi.org/10.1016/j.aeue.2011.05.002) 16Heinzelman W., Chandrakasan A., Balakrishnan H.: ' Energy-efficient communication protocol for wireless microsensor networks'. Proc. of the 33rd Int. Conf. on System Science (HICSS'00), Hawaii, USA, 2000, pp. 1– 10 17Singh B., Lobiyal D.K.: 'A novel energy-aware cluster head selection based on particle swarm optimization for wireless sensor networks', Hum.-Centric Comput. Inf. Sci., 2012, 2, (13) 18Attea B.A., Khalil E.A.: 'A new evolutionary based routing protocol for clustered heterogeneous wireless sensor networks', Appl. Soft Comput., 2012, 12, (7), pp. 1950– 1957 (doi: https://doi.org/10.1016/j.asoc.2011.04.007) 19Hoang D.C., Yadav P., Kumar R. et al.: 'Real-time implementation of a harmony search algorithm-based clustering protocol for energy-efficient wireless sensor networks', IEEE Trans. Ind. Inf., 2014, 10, (1), pp. 774– 783 (doi: https://doi.org/10.1109/TII.2013.2273739) 20Karaboga D., Okdem S., Ozturk C.: 'Cluster based wireless sensor network routing using artificial bee colony algorithm', Wirel. Netw., 2012, 18, (7), pp. 847– 860 (doi: https://doi.org/10.1007/s11276-012-0438-z) 21Chen R.-C., Chang W.-L., Shieh C.-F. et al.: ' Using hybrid artificial bee colony algorithm to extend wireless sensor network lifetime'. Proc. of Third Int. Conf. on Innovations in Bio-Inspired Computing and Applications, 2012 22Yan F., Yeung A.K.H., Joseph A.C. et al.: 'Degree-energy-based local random routing strategies for sensor networks', Commun. Nonlinear Sci. Numer. Simul., 2015, 20, (1), pp. 250– 262 (doi: https://doi.org/10.1016/j.cnsns.2014.04.017) 23Han Z., Wu J., Zhang J. et al.: 'A general self-organized tree-based energy-balance routing protocol for wireless sensor network', IEEE Trans. Nucl. Sci., 2014, 61, (2), pp. 732– 740 (doi: https://doi.org/10.1109/TNS.2014.2309351) 24Radi M., Dezfouli B., Bakar K.A. et al.: 'Multipath routing in wireless sensor networks: survey and research challenges', Sensors, 2012, 12, (1), pp. 650– 685 (doi: https://doi.org/10.3390/s120100650) 25Ganesan D., Govindan R., Shenker S. et al.: 'Highly-resilient, energy-efficient multipath routing in wireless sensor networks', ACM Mob. Comput. Commun. Rev. (MC2R), 2001, 5, pp. 11– 25 (doi: https://doi.org/10.1145/509506.509514) 26De S., Qiao C., Wu H.: 'Meshed multipath routing with selective forwarding: an efficient strategy in wireless sensor networks', Wirel. Sens. Netw., 2003, 43, pp. 481– 497 27Sharma S., Jena S.K.: 'Cluster based multipath routing protocol for wireless sensor networks', Newsl., ACM SIGCOMM Comput. Commun. Rev., 2015, 45, (2), pp. 14– 20 (doi: https://doi.org/10.1145/2766330.2766333) 28Jin R.-C., Gao T., Song J.-Y. et al.: 'Passive cluster-based multipath routing protocol for wireless sensor networks', Wirel. Netw., 2013, 19, (8), pp. 1851– 1866 (doi: https://doi.org/10.1007/s11276-013-0570-4) 29Zhong C., Mo Y., Zhao J. et al.: ' Secure clustering and reliable multi-path route discovering in wireless sensor networks'. Proc. 2014 Sixth Int. Symp. on Parallel Architectures, Algorithms and Programming (PAAP), 2014, pp. 130– 134 30Heinzelman W.B., Chandrakasan A.P., Balakrishnan H.: 'An application-specific protocol architecture for wireless microsensor networks', IEEE Trans. Wirel. Commun., 2002, 1, (4), pp. 660– 670 (doi: https://doi.org/10.1109/TWC.2002.804190) 31Grahama M.A., Mukherjee A., Chakraborti S.: 'Distribution-free exponentially weighted moving average control charts for monitoring unknown location', Comput. Stat. Data Anal., 2012, 56, (8), pp. 2539– 2561 (doi: https://doi.org/10.1016/j.csda.2012.02.010) 32Dorigo M., Gambardella L.M.: 'Ant colony system: a cooperative learning approach to the traveling salesman problem', IEEE Trans. Evol. Comput., 1997, 1, (1), pp. 53– 66 (doi: https://doi.org/10.1109/4235.585892) 33Okdem S.: 'A real-time noise resilient data link layer mechanism for unslotted IEEE 802.15.4 networks', Int. J. Commun. Syst., 2015 34Wang F., Li D., Zhao Y.: 'Analysis of CSMA/CA in IEEE 802.15.4', IET Commun., 2011, 5, (15), pp. 2187– 2195 (doi: https://doi.org/10.1049/iet-com.2010.1007) Citing Literature Volume11, Issue4March 2017Pages 522-530 FiguresReferencesRelatedInformation
Referência(s)