Elephant swarm optimization for wireless sensor networks –a cross layer mechanism
Upcoming SlideShare
Loading in...5
×
 

Elephant swarm optimization for wireless sensor networks –a cross layer mechanism

on

  • 498 views

 

Statistics

Views

Total Views
498
Views on SlideShare
498
Embed Views
0

Actions

Likes
0
Downloads
1
Comments
0

0 Embeds 0

No embeds

Accessibility

Categories

Upload Details

Uploaded via as Adobe PDF

Usage Rights

© All Rights Reserved

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
    Processing…
Post Comment
Edit your comment

    Elephant swarm optimization for wireless sensor networks –a cross layer mechanism Elephant swarm optimization for wireless sensor networks –a cross layer mechanism Document Transcript

    • INTERNATIONALComputer Engineering and Technology ENGINEERING International Journal of JOURNAL OF COMPUTER (IJCET), ISSN 0976- 6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 2, March – April (2013), © IAEME & TECHNOLOGY (IJCET)ISSN 0976 – 6367(Print)ISSN 0976 – 6375(Online) IJCETVolume 4, Issue 2, March – April (2013), pp. 45-60© IAEME: www.iaeme.com/ijcet.aspJournal Impact Factor (2013): 6.1302 (Calculated by GISI) ©IAEMEwww.jifactor.com ELEPHANT SWARM OPTIMIZATION FOR WIRELESS SENSOR NETWORKS –A CROSS LAYER MECHANISM Chandramouli.H1, Dr. Somashekhar C Desai2, K S Jagadeesh3 and Kashyap D Dhruve4 1 Research Scholar, JJT University, Jhunjhunu, Rajasthan, India. 2 Professor, BLDEA College of Engineering, Bijapur, India. 3 Research Scholar, JJT University, Jhunjhunu, Rajasthan, India. 4 Technical Director, Planet-I Technologies, Bangalore, India. ABSTRACT The dominant quality of service (ܳ‫ )ܵ݋‬oriented parameters in wireless sensing networks (ܹܵܰ‫ )ݏ‬are network lifetime, network throughput, overheads and of course the depicting the quality of a particular network and even in ܹܵܰ‫ ݏ‬power is considered as the overall network efficiency. Power is considered as one of the most important factors in most influensive parameter. The life time of any ܹܵܰ‫ ݏ‬network is in proportion with the power level allied with the network components. Therefore the enhancement of the lifetime of wireless sensor networks has ignited the entire communication research society and scientific society to develop a robust system architecture that can effectively optimize the parameters like network lifetime, throughput etc. Now days a number of algorithms and protocols are being developed based on the nature of evolution and its characteristics to develop real time protocols for certain optimization problems. Considering the behavior of communication protocol for enhancing the network ܳ‫ ܵ݋‬with enhanced network lifetime and elephant swarm as an inspiration the researcher of this research paper has developed a robust throughput. The authors of this research manuscript draw inspiration from the behavior of large elephant swarms and incorporate their behavior into wireless sensor networks. The characterizing behavior of elephant swarm has been incorporated using the robust cross layer approach. The elephant swarm optimization technique being discussed in this research paper and balanced ܶ‫ ܥܣܯ ܣܯܦ‬scheduling so as to achieve a cumulative enhanced network facilitates a robust as well as optimized routing technique, adaptive radio link optimization performance. In this research work the proposed elephant swarm optimization has been compared with another robust evolutionary computing based optimization technique called as Particle swarm Optimization (ܱܲܵ). The experimental study presented proves that the Elephant Swarm Optimization technique enhances the network life time by about 26.8%. 45
    • International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 2, March – April (2013), © IAEMEKeywords: Elephant Swarm, Optimization, Network Lifetime, Cross-layer design, energyefficiency, resource allocation, sensor networks, Evolutionary Algorithm, Node Decay,Communication Overhead, Particle Swarm Optimization, Active Node Ratio1. INTRODUCTION Consider a network topology of wireless sensor network deployed over a particulargeographical region. The WSNs sensor nodes are considered to be possessing homogenousenergy properties and are battery operated which is the case most often than not. In order toachieve higher data transmission rate and uniform load distribution the sensor distributionover the geographical region is considered to be dense. Owing to dense deployments anumber of links are established stimulate interference amongst the sensor nodes that isrequired to be minimized so as to achieve optimum network performance in terms of networkthroughput and higher lifetime. This paper introduces an Elephant Based SwarmOptimization model so as to enhance network life time. Cross layer approach and systemarchitecture has been adopted to for incorporating the elephant swarm optimization features. Elephants are social animals [1] and are characterized [2] by the nature of theiradvanced intelligence. The existence of elephants is often found in a “fluid fission-fusion”social environment [3]. Elephants are generally characterized by their good memory, theirnature to coexist, group leading and survive within a “clan” [4] (swarm of more than athousand elephants) socially formulated during testing times like migration and when theresources are scare. The exhibition of the unselfish behavior enables them to grow and is thesecret of their longevity. Keeping progress and survivability in mind the older elephantsdisassociate from the “clan”. Elephants by nature are protective of their younger generation.Elephants swarm communicates using varied advanced techniques which include acousticcommunication, chemical communication, visual communication and tactile communication[5] [6]. Memory of elephant is considered as one of the most important characteristics tosurvive and lead the clan. Their memory empowers them with recognition, identification andproblem solving scenarios [4]. All these features exhibited have influenced the authors toincorporate such behavior in wireless sensor networks to improve network performance. The implementation of elephant swarm model is in fact a complex procedure and inorder to realize such behaviors in wireless sensor networks the authors have proposed asystem architecture that adopts a cross layer approach to incorporate the elephant swarmmodel. The network optimization is a must factor for adopting at the Routing Layer, MACLayer and the Radio Layer of the wireless sensor node. This research paper introduces anenhanced and robust cross layer approach to incorporate the elephant swarm optimizationtechnique which is compared with the popular Particle Swarm Optimization technique and itsefficiency is proved in the latter section of this paper. The remaining manuscript is organized as follows. Section two discusses a briefliterature study conducted during the course of the research work presented here. The systemmodelling and the elephant swarm optimization technique using a cross layer approach isdiscussed in the third section of this paper. The experimental study conducted is described inthe penultimate section of this paper. The conclusions drawn and the future work is presentedin the last section of this paper. 46
    • International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 2, March – April (2013), © IAEME2. LITERATURE SURVEY The requirement and hence the enhancement of the existing techniques and coming upwith an optimized solution of network life time as well as throughput enhancement hasignited the scientific society to come up with an optimum solution and approach so that theneed can be met. In philosophical and as a knowledge strengthening factor the literaturereview is considered as a fundamental need of research work. The author of this researchwork has surveyed some algorithms. Few of those are as follows: In [7] investigated the cross layer survivable link mapping when the traffic layers areunambiguously desired and survivability is must. In this work a forbidden link matrix isidentified the masking region of the network for implementing in such conditions wheresome physical links are reserved exclusively for a designated service, mainly for the contextof providing multiple levels of differentiation on the network use. The masking upshot is thenestimated on two metrics using two sensible approaches in a real-world network, depictingthat both effectiveness and expediency can be obtained. The literature [8] the researcher proposed a route discovery and congestion handlingmechanism that employs a cross layer model including a potential role in congestiondetection and its regularization. The limitation of the proposed technique was its confineddata rate. Considering the dominant network parameters like deploy, energy consumption,that employs a cross layer ‫ ܥܣܯ‬protocol for wireless network. This work employs theexpansibility, flexibility and error tolerance Jin, Lizhong et al [9] presented a research worksplitting of ‫ ܥܣܯ‬layer and of course it performed well, but considering the higher data rate In literature [10] proposed a new cross layer-based ‫ ܥܣܯ‬protocol stated as ‫.ܥܣܯܮܥ‬transmission this system was found to be ineffective even having more error prone.In this proposed cross layered ‫ ܥܣܯ‬technique, the communications among ‫ ,ܥܣܯ‬Routingand Physical layers are fully exploited so as to minimize the energy consumption and multi-carrier-sensing technology is applied at the ܲ‫ ܻܪ‬layer so as to sense the traffic load andhop delay of the data delivery for wireless sensor networks. In precise, in that approach thenecessarily initiates the neighbour nodes in multi hops so that the data transmission can berealized over multihop. Similarly, by implementing the routing layer information, thedeveloped cross layered MAC facilitates the receiver of the ascending hop on the path ofrouting that has to be effectively waken up and ultimately it results into the potentialreduction in energy consumption.techniques at ‫ ܥܣܯ‬layer were considered In literature [11] a number of fundamental cross layered resource allocation for fading channel. This research worknetwork layer, ‫ ܥܣܯ‬layer quality and physical layer as performance.emphasizes on characterization of fundamental performance limits while considering the Hang Su [12] proposes the cross layer architecture based an opportunistic ‫ܥܣܯ‬protocol that integrates the spectrum sensing at ܲ‫ ܻܪ‬layer and packet scheduling at the ‫ܥܣܯ‬layer. In their proposal the secondary user is equipped with two transceivers where one istuned for dedicated control channel while another one is designed particularly for cognitiveradio that can effectively use the idle radio. They propose two shared channel spectrum-policy so as to assist the ‫ ܥܣܯ‬protocols detect the availability of leftover channels. Thissensing approach, named as the random sensing policy and the negotiation-based sensingfrequency and thus the other ܳ‫ ܵ݋‬parameters are not being considered.technique has a great potential but the emphasis has been made on the efficient use of leftover 47
    • International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 2, March – April (2013), © IAEME A research work [13] describes the fundamental concept of sensor networks whichhas been made viable by the convergence of micro electro-mechanical systems technology,wireless communications and digital electronics. In their work initially the potential sensornetworks are explored and then the dominant factors influencing the system architecture ofnetwork is obtained and in the later stage the communication architecture was outlined andthe algorithms were developed for different layers of the network for system optimization. Asthis proposal brought certain positive results but was lacking the optimized output and havinga lot of vacuum for further development. The researcher in [14] developed a recommender system, employing a particle swarmoptimization (ܱܲܵ) algorithm for learning the personal preferences of users and facilitates thetailored solutions. The system being used in this research was based on collaborative filteringapproach, building up profiles of users and then using an algorithm to find profiles similar tothe current user. To overcome the problem of sparse or implicated data they utilizedmatching and finally the ܱܲܵ was used to optimize the results. That system was found to bestochastic and heuristic-based based models to speed up and improve the quality of profileoutperforming genetic algorithm concept but the system could not play a vital role in higherdata rate with cross layer architecture and especially for heterogeneous type of network.3. ELEPHANT SWARM OPTIMIZATION – CROSS LAYER ARCHITECTURE3.1 SYSTEM MODELING The system modeling section represents the approach and techniques beingimplemented so as to realize the elephant swarm optimization for wireless sensor networks isdiscussed. Let us consider a‫ ݓ‬wireless sensor nodes represented by a setܹ which constitute astatic network defined asܹ ൌ ሼ‫ݓ‬ଵ , ‫ݓ‬ଶ , ‫ݓ‬ଷ , … . ‫ݓ‬௪ ሽ … … . . ሺ1ሻtwo nodes ‫ݓ‬ଵ ‫ܹ א‬and ‫ݓ‬ଶ ‫ ,ܹ א‬a relatively high transmission power allocation scheme is In the considered networkܹ , the wireless communication links that exist betweenconsidered . The high power allocation scheme causes the higher power consumption thatultimately results into numerous interferences situation between other nodes as well asdegraded network life time and hence poor efficiency. The communication channel beingconsidered over the links is nothing but Additive White Gaussian Noise (‫ )ܰܩܹܣ‬channelmodel has been assumed. If the signal to noise ratio ሺܵ‫ܴܰܫ‬ሻ of a communication link ishaving confined noise power level. Here, one more factor called deterministic path lossrepresented by ߛ then the maximum data rate supported ሺ݉௥ ሻper unit bandwidth is definedas݉௥ ൌ ݈‫ ݃݋‬൫1 ൅ ሺ‫ ܤ‬ൈ ߛሻ൯ … … . . ሺ2ሻ െ1 · 5Where‫ ܤ‬ൌ … … . . ሺ3ሻ ݈‫݃݋‬ሺ5‫ ܴܧܤ‬ሻ 48
    • International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 2, March – April (2013), © IAEMEconstellation size for the‫ ܯܣܳܯ‬is൒ 4 and varies with time over a considered link [15].The This considered model can be realized using modulation schemes like‫ .ܯܣܳܯ‬Themodel assumes a ܶ‫ܣܯܦ‬scheduling system of communication between the nodes. The modelconsidered assumes that there exists ܰ௧ time slots for the medium access control layer (‫)ܥܣܯ‬Let us consider that a particular node ‫ݓ‬௪ ‫ܹ א‬transmits at a power level P୲ then the powerand a unique transmission mode is applicable per slot.consumption of the amplifier is defined asሺ1 ൅ αሻP୲ … … . . ሺ4ሻWhere α is the efficiency of power amplifier and ߙ ൐ 0 to achieve the desired signalamplification.A homogenous sensor network model is considered݅. ݁. ‫ݓ׊‬௪ ‫ߙ ׷ ܹ א‬ଵ ൌ ߙଶ ൌ ‫ ڮ‬ൌ ߙ௪ .The directed graph that represents the network ܹunder consideration, is defined as‫ܦ‬௚ ൌ ሺ ܹ ‫ܮ‬ሻ … … . . ሺ5ሻWhere L indicates set of directed links.Letࣛ ‫| ࣬ א‬ௐ|ൈ|୐| indicates the incidence matrix of the graph ‫ܦ‬௚ then we can state that െ1 ‫݂ܫ‬w௪ is the receiver of link ℓࣛሺw௪ , ℓሻ ൌ ൝ 0 ݅݊‫ ݏ݁ݏܽܿݎ݄݁ݐ݋‬ൡ … … . . ሺ6ሻ ൅1 ‫݂ܫ‬w௪ is the transmitter of link ℓWe present an expressionࣛ ൌ ࣛା െ ࣛି … … . . ሺ7ሻSuch that ࣛ ା ሺआ, ℓሻ, ࣛ ି ሺआ, ℓሻ ൌ 0, and ࣛ ା , ࣛ ି and have the entries of 0 and 1.As discussed earlier ܰ௧ is the number of time slots in individual frame of the periodicschedule. ‫ܮ‬௡೟ represents the set of link scheduled. These are allowed to transmit during timeslot defined as݊௧ ‫ א‬ሼ1, ‫ܰ , ڮ ڮ ڮ‬௧ ሽ … … . . ሺ8ሻܲ௟ ೟ and݉௥ ௟ ೟ represents the power of transmission and per unit bandwidth rate respectively ௡ ௡over link ݈and ݊௧ slots.The vectors of the time slot ݊௧ are ݉௥ ௡೟ and ܲ௡೟ ‫ܲ. |୐| ࣬ א‬ℓࣾࣵई is theanalogous vector isܲ ࣾࣵई ‫ . |୐| ࣬ א‬The vectors 1ऄ ሺܲ ௡೟ ሻid defined asmaximum limit of allowable transmission power for the node which belongs to link݈. The 1 ݂݅ሺሺeା ሻ் ൈ ܲ௡೟ ሻ ൐ 0൫1௧ ሺܲ௡೟ ሻ൯௪ ൌ ൜ आ ൠ … … . . ሺ9ሻ ೢ 0 ‫ݏ݁ݏܽܿݎ݄݁ݐ݋݊ܫ‬Where ሺeା ሻ் is the νऄࣺ row of the matricesࣛା . Also ൫1௧ ሺܲ ௡೟ ሻ൯௪ ‫| ࣬ א‬୛| आ ೢ 49
    • International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 2, March – April (2013), © IAEMEThe vector 1௠ೝ ሺܲ ௡೟ ሻ is defined asቀ1௠ೝ ሺܲ ௡೟ ሻቁ ൌ ൜1 ݂݅ሺሺeି ሻ் ൈ ܲ ௡೟ ሻ ൐ 0ൠ आ … … . . ሺ10ሻ ௪ೢ 0 ‫ݏ݁ݏܽܿݎ݄݁ݐ݋݊ܫ‬Where ሺeି ሻ் is the νऄࣺ row of the matricesࣛି .Also ቀ1௠ೝ ሺܲ ௡೟ ሻቁ आ ‫| ࣬ א‬୛| ௪ೢThe initial homogenous energy of all the nodes ‫ݓ‬௪ ‫ ܹ א‬defined asࣟ௪ೢ and the energy ࣟ ‫| ࣬ א‬୛|Let ܲ௧௖௢௡ represents power consumption of transmitter and ܲ௥௖௢௡ represents the powerconsumed by each node ‫ݓ‬௪ ‫ܹ א‬is ൑ ࣟ௪ೢconsumption of the receiver and is assumed to be homogenous for all the nodes. Therepresented as ࣭௪ೢ . It can be stated that ࣭ ‫| ࣬ א‬୛| represents a vector which constituteLet the sensing events that are induced in the network induce an information generation rateof࣭௪ೢ .The data aggregated at the sink is defined as࣭࣭ࣻࣿࣽ ൌ െ ∑௪ೢ ‫א‬ௐ,௪ೢ ஷ࣭ࣻࣿࣽ ࣭௪ೢ … … . . ሺ11ሻ‫ܦ‬௚ ൌ ࣬ |୐|ൈ|୐| … … . . ሺ12ሻThe link gain matrix of the wireless sensor network considered is defined asThe power from the transmitter of the ࣥ ௧௛ link to the receiving node on link݈is represented as‫ܦ‬௚ and ࣨ଴ represents the total noise power over the operational bandwidth. ௟ࣥThe ܶ௪ೢ represent the network lifetime when a percentage of nodes ‫ %ݓ‬runs out of energy.This is a common criterion considered by researchers to evaluate their proposed algorithms[ref].The maximum data rate supported for transmission over a particular Link݈ ‫ܮ א‬is defined as൫1 ൅ ሺࣥ ൈ ݈‫݃݋‬ሺܵ‫ܴܰܫ‬ሻሻ൯ … … . . ሺ13ሻ3.2 ELEPHANT SWARM OPTIMIZATION OBJECTIVE A cross layer approach is adopted to enhance the network lifetime of the wirelesssensor network. Elephants are social animals and are said to possess strong memory of theevents that occurThe problem of optimizing or maximizing the life span of the network can be presented as aࣾࣵई. ሺ࣮ ሻfunction defined as follows ࣿՂऄ݅. ݁. ࣛ ሺ݉௥ ଵ ൅ ‫ ڮ ڮ ڮ‬൅ ݉௥ ே೟ ሻ ൌ ࣭ … … . . ሺ14ሻ ଵ ே೟ 50
    • International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 2, March – April (2013), © IAEME ࣿ ஽೒ ࣪ℓ݈‫ ݃݋‬൬1 ൅ ࣥ ∑ ൰ ൒ ݉௥ ௟ ೟ … … . . ሺ15ሻ ೗೗ ࣿ ࣥಯℓ ஽೒ ೗ೖ ࣪ࣽ ା ࣨ ࣿ బThe maximization function ࣾࣵई. ሺ࣮ ሻ can be defined as ࣿՂऄ ே೟࣮ ෍ ቀሺ1 ൅ ߙሻࣛା ࣪ ࣿ೟ ൅ ܲ௧௖௢௡ 1ऄ ሺ࣪ ࣿ೟ ሻ ൅ ܲ 1௠ೝ ሺ࣪ ࣿ೟ ሻቁ ‫݉ࣟ ع‬௥ ࣿ೟ ‫ࣿ ࣪ ع 0,0 غ‬೟ ‫ࣵࣾ ࣪ ع‬ई … … . . ሺ16ሻ ࣿՂऄ ܰ௧ ௥௖௢௡ ࣿ೟ ୀଵFor all time slotsࣿ௧ ൌ 1, ‫ܰ , ڮ ڮ‬௧ and݈ ‫ ,ܮ א‬the constituting variables are࣮ ,݉௥ ௟ ೟ ,࣪ℓ ೟ , for ࣿ ࣿ ࣿՂऄa set ࣿ௧ ‫ א‬ሼ1, ‫ܰ , ڮ ڮ‬௧ ሽ, ݈ ‫.ܮ א‬Let us define a variable ܱsuch thatܱ ൌ ሺ࣮ ሻିଵ ࣿՂऄ … … . . ሺ17ሻࣾ݅݊. ሺܱሻThe elephant swarm optimization is applied to attain minimized function defined as݅. ݁. ࣛ ሺ݉௥ ଵ ൅ ‫ ڮ ڮ ڮ‬൅ ݉௥ ே೟ ሻ ൌ ሺ࣭ ൈ ܰ௧ ሻ … … . . ሺ18ሻ ‫ܦ‬௚ ࣪ℓ ࣿ೟݈‫ ݃݋‬൭1 ൅ ൱ ൒ ݉௥ ௟ ೟ … … . . ሺ19ሻ ௟௟ ࣿ ∑ࣥஷℓ ‫ܦ‬௚ ࣪ࣽ ೟ ࣿ ൅ ࣨ଴ ௟௞The minimization function or the elephant swarm optimization objective ࣾ݅݊. ሺܱሻ can bedefined as ே೟෍ ቀሺ1 ൅ ߙሻࣛା ࣪ ࣿ೟ ൅ ܲ௧௖௢௡ 1ऄ ሺ࣪ ࣿ೟ ሻ ൅ ܲ௥௖௢௡ 1௠ೝ ሺ࣪ ࣿ೟ ሻቁ ‫ܱܰࣟ ع‬௧ … … . . ሺ20ሻࣿ೟ ୀଵThe model presented here considers ܶ‫ܣܯܦ‬based ‫ܥܣܯ‬systems the minimization function canbe defined asࣾ݅݊. ሺܱሻ … … . . ሺ21ሻ ை݅. ݁. ෍ ሺܽ௟ ሻ െ ෍ ሺܽ௟ ሻ ൌ ܰ௧ ࣭௪ೢ … … . . ሺ22ሻ ௟‫்א‬௫ሺ௪ೢ ሻ ௟‫א‬ோ௫ሺ௪ೢ ሻ ‫ܦ‬௚ ࣪௟ࣾࣵईܽ௟ െ ࣿ௧ ௟ ݈‫ ݃݋‬ቆ1 െ ௟௟ ቇ൑0 … … . . ሺ23ሻ ܰ଴ ೌℓ ܲ௧௖௢௡ ෍ ߞࣿ௧ ௟ ቆՂ ࣿ೟೗ െ 1 ൅ ቇ ൅ ෍ ܲ௥௖௢௡ ࣿ௧ ௟ ൑ ܱܰ௧ ࣟ௪ೢ … … . . ሺ24ሻ ߞ௟‫்א‬௫ሺ௪ೢ ሻ ௟‫א‬ோ௫ሺ௪ೢ ሻ 51
    • International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 2, March – April (2013), © IAEME෍ ࣿ௧ ௟ ൑ ܰ௧ , ܽ௟ , ࣿ௧ ௟ ൒ 0, ࣿ௧ ௟ ‫ א‬ሼ0, ‫ܰ , ڮ ڮ‬௧ ሽ … … . . ሺ25ሻ௟‫א‬௅Where݈represents the link, the number of slots assigned on݈isࣿ௧ ௟ . ܶ‫ ݔ‬ሺ‫ݓ‬௪ ሻis the set oftransmitting links and ܴ‫ݔ‬ሺ‫ݓ‬௪ ሻis the receiving links of the sensor node ‫ݓ‬௪ ‫.ܹ א‬The variableߞis defined as follows ܰ଴ ሺ1 ൅ ߙሻߞ ൌ … … . . ሺ26ሻ ‫ܦ‬௚ ௟௟And ܽ௟ is defined asܽ௟ ൌ ݉௥ ௟ ൈ ࣿ௧ ௟ … … . . ሺ27ሻThe transmission power the݈௧௛ link represented as ࣪௟ is defined as࣪௟ ൌ ஽ బ ሺՂ ୫౨ ℓ െ 1ሻ … … . . ሺ28ሻ ୒ ೒ ೗೗It must be noticed that the power of transmission over a network link ݈ is presented as ܰ଴ ௠ ௟࣪௟ ൌ ሺՂ ೝ െ 1ሻ … … . . ሺ29ሻ ‫ܦ‬௚ ௟௟3.3 PHASED APPROACH TO REALIZE ELEPHANT SWARM BEHAVIOUR The presented section of this paper elaborates the elephant swarm optimizationalgorithm for routing,ܶ‫ ܥܣܯܣܯܦ‬scheduling and advanced radio layer control techniques.the network links. The elephant swarm optimization enables simultaneous ܶ‫ ܣܯܦ‬schedulingThe elephant swam optimization is applied taking into account unconstrained scheduling onof the sensing data on the interfering wireless communication links in the current consideredpower consumption and ܶ‫ܥܣܯܣܯܦ‬schedule to enhance the considered network lifetime.scheduling time slot. The elephant swarm optimization iterates to obtain an optimal routing,The elephant model is adopted to solve optimization objective ࣾ݅݊. ሺܱሻ defined in theformer section of this paper.Let us consider a ܶ‫ܣܯܦ‬link schedule of data defined as ‫ܮ‬௡೟ where݊௧ ‫ א‬ሼ1, ‫ܰ , ڮ ڮ ڮ‬௧ ሽ. Therate of transmission that can be supported over a link݈ ‫ܮ א‬can be expresses as based on ‫ܦ‬௚ ࣪௟ ೟ ࣿapproximations is defined as݉௥ ௟ ൑ log ൭ ൱ … … . . ሺ30ሻ ࣿ೟ ௟௟ ∑௞ஷℓ,௞‫א‬௅೙೟ ‫ܦ‬௚ ࣪௞ࣿ೟ ൅ ࣨ଴ ௟௞If the ܵ‫ܴܰܫ‬of a link݈ ‫ܮ א‬is ߛthen the minimum transmission rate is defined as follows݈‫݃݋‬ሺߛሻ … … . . ሺ31ሻ݉௥ ௟ ೟ are said to be a part of the ࣾ݅݊. ሺܱሻfunction optimization set. ࣿThe elephant swarm optimization results arising based on the above approximations for 52
    • International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 2, March – April (2013), © IAEMELet us define a variableࣤ௟ ൌ log൫࣪௟ ೟ ൯ … … . . ሺ32ሻ ࣿ೟ ࣿThen the above equation for the maximum transmission rate optimization ݉௥ ௟ ೟ can be ࣿdefined as can be ܰ଴ ௠ೝ ࣿ೟ ‫ܦ‬௚݈‫ ݃݋‬൭ Ղ ೗ െ ࣤ௟ࣿ ൅ ෍ Ղ ௠ೝ ℓ ൅ ࣤࣽ ೟ െ ࣤℓ ೟ ൱ ൑ 0 … … . . ሺ33ሻ ࣿ೟ ௟ࣽ ࣿ ࣿ ‫ܦ‬௚ ‫ܦ‬௚ ௟௟ ࣥ‫א‬௅ࣿ ,ࣥஷ௅ ௟௟Based on the above arguments the elephant swarm optimization objective ࣾ݅݊. ሺܱሻcan beexpressed asࣛ ሺ݉௥ଵ ൅ ‫ ڮ ڮ ڮ‬൅ ݉௥ ே೟ ሻ ൌ ሺ࣭ ൈ ܰ௧ ሻ … … . . ሺ34ሻ ܰ଴ ௠ೝ ࣿ೟ ‫ܦ‬௚݈‫ ݃݋‬൭ Ղ ೗ െ ࣤ௟ࣿ ൅ ෍ Ղ ௠ೝ ℓ ൅ ࣤࣽ ೟ െ ࣤℓ ೟ ൱ ൑ 0 , ‫ . . … … ܮ א ݈׊‬ሺ35ሻ ࣿ೟ ௟ࣽ ࣿ ࣿ ‫ܦ‬௚ ‫ܦ‬௚ ௟௟ ࣥ‫א‬௅ࣿ ,ࣥஷ௅ ௟௟ ே೟෍ቌ ෍ ൬ሺ1 ൅ ߙሻՂ ࣤ೗ ൅ ܲ௧௖௢௡ ൰ ൅ ෍ ሺܲ௥௖௢௡ ሻቍ ൑ ܱܰ௧ ࣟ௪ೢ … … . . ሺ36ሻ ࣿ೟ࣿ೟ ୀଵ ௟‫்א‬௫ሺ௪ೢ ሻ‫ת‬௅೙೟ ௟‫א‬ோ௫ሺ௪ೢ ሻ‫ת‬௅೙೟݉ ௥ ࣿ೟ ൒ 0 The above defined elephant swarm optimization is applicable providedࣤ௟ ೟ ൑ logሺ࣪௟௠௔௫ ሻ ݈ࣿ ‫ܮ א‬௡೟ and݊௧ ൌ ሼ1, ‫ܰ , ڮ ڮ ڮ‬௧ ሽ … … . . ሺ37ሻIn other terms the elephant swarm optimization is applicable if the links have a ܵ‫ܴܰܫ‬greaterThe ܶ‫ ܥܣܯܣܯܦ‬scheduling over all the links is not adopted as the power consumption wouldthan unity.exponentially increase. The elephant swarm optimization is applied on all the ܶ‫ ܣܯܦ‬linksscheduled‫ܮ‬௡೟ . The computational complexity of optimization under such circumstances canbe defined as‫ܥ‬஼௢௠௣௟௘௫ ൌ 2ሺ௅ൈே೟ ሻ … … . . ሺ38ሻFrom the above equation it is evident that the ܶ‫ ܥܣܯܣܯܦ‬optimization is computationallynetworks) and the ܶ‫ ܣܯܦ‬slot value increases. The computation complexity of the elephantheavy and increases exponentially as the links of the sensor nodes increase (i.e. for denseswarm optimization can be reduced if the number of ܶ‫ ܥܣܯܣܯܦ‬slots are doubled to 2ܰ௧ .The two fold increase in the number of time slots enables achieving lower powerconsumption as the sensor nodes have numerous slot options and sleep induction is effective.Furthermore in the case of high sensing activity leading to greater data transmissions, the datato the sink is scheduled using multiple TDMA slots to enable energy conservation andaccurate data aggregation. 53
    • International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 2, March – April (2013), © IAEMEThe elephant swarm optimization model can be summarized in the form of the algorithmgiven below realized through multiple phases described below.Phase A:Initialize the schedule ‫ܮ‬௡೟ based on the data࣭௪ೢ . The ‫ܮ‬௡೟ is initialized such thatlink݈ ‫א‬‫ܮ‬௡೟ ‫݊׌݈׊‬௧ ‫ܰ א‬௧ .݅. ݁. the schedule is constructed in a manner such that all the links ݈ ‫ܮ א‬௡೟ areprovided at least a slot in ܰ௧Phase B:In this phase the following equation is solved ே೟෍ቌ ෍ ൬ሺ1 ൅ ߙ ሻՂ ࣤ೗ ൅ ܲ௧௖௢௡ ൰ ൅ ෍ ሺܲ௥௖௢௡ ሻቍ … … . . ሺ39ሻ ࣿ೟ࣿ೟ ୀଵ ௟‫்א‬௫ሺ௪ೢ ሻ‫ת‬௅೙೟ ௟‫א‬ோ௫ሺ௪ೢ ሻ‫ת‬௅೙೟If the results obtained on solving are not suitable . ݅. ݁. ൐ ܱܰ௧ ࣟ௪ೢ then the optimization isnot possible. If the solutions satisfy the condition ൑ ܱܰ௧ ࣟ௪ೢ then elephant swarm routeoptimization and radio layer optimizations are carried out to support the requiredtransmission rate.Phase C:Evaluate all the links ݈ ‫ܮ א‬௡೟ and retain the links if the following equation is satisfied. ‫ܦ‬௚ ࣪௟ ࣿ೟ ௟௟ ൐ ߛ௢ … … . . ሺ40ሻ∑௞ஷℓ,௞‫א‬௅೙೟ ‫ܦ‬௚ ࣪௞ࣿ೟ ൅ ࣨ଴ ௟௞This phase eliminates all the links whose ܵ‫ ܴܰܫ‬is less than unity and retaining the linkshaving an acceptableܵ‫.ܴܰܫ‬Phase D: ሸCompute݈ using the following equation ே೟ሸ ൌ ݈ெ௔௫ ቌ ෍ ൫࣪ ࣿ೟ ൯ቍ݈ ஼௥௡௧ … … . . ሺ41ሻ ௟ ࣿ೟ ୀଵ ഺCompute ࣿ௧ defined asഺࣿ௧ ൌ ࣿ௧ ஼௥௡௧ ቌ ෍ ‫ܦ‬௚ ሸ ࣪௞ ೟ ൅ ࣨ଴ ቍ … … . . ሺ42ሻ ࣿ ெ௜௡ ௟௞ ሸ ௞ஷ௟,௞‫א‬௅೙೟ 54
    • International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 2, March – April (2013), © IAEMEUsing the above definitions we can obtain the new the ܶ‫ ܥܣܯܣܯܦ‬layer schedule ഺ ሸrepresented as ‫ࣿܮ‬೟ and݈ ‫ࣿܮ א‬೟ . If the optimized ܶ‫ࣿܮܥܣܯܣܯܦ‬೟ schedule is equitant to the ഺ ഺexisting or previous schedule then no further optimization is possible. If the optimized ‫ࣿܮ‬೟ is ഺ‫ࣿܮ‬೟ enables to identify the maximum power utilization link݈ so we can schedule it thenot similar to the current and previous MAC layer schedule the new schedule is adopted. ഺadditional slots available thus aiding energy conservation.Phase E:In the last phase of the elephant swarm optimization algorithm the optimal solution achievedusing a cross layer approach is verified using the following definition ‫ܦ‬௚ ࣪௟ ࣿ೟൭ ௟௟ ൱ ൒ 1.0 ‫ܮ א ݈׊‬௡೟ , ݊௧ ൌ ሼ1, ‫ܰ , ڮ ڮ ڮ‬௧ ሽ … … . . ሺ43ሻ ∑௞ஷℓ,௞‫א‬௅೙೟ ‫ܦ‬௚ ࣪௞ࣿ೟ ൅ ࣨ଴ ௟௞If the solution does not satisfy the above equation then no optimization is possible owing tocurrent network dependent reasons. If optimal solution is obtained and incorporated networkperformance in terms of data aggregation, improved data rates and higher network lifetimes. The elephant swarm optimization is realized using a cross layer design approach toenhance network lifetime. The efficiency and the performance measure of this optimizationtechnique is discussed in the subsequent section of this paper.4. EXPERIMENTAL STUDY AND COMPARISONS This section represents and elaborates the experimental set up with the resultsobtained and their respective analysis with depicting their significance for the proposedresearch work. This section discusses the experimental study conducted to compare theelephant swarm optimization algorithm introduced in this paper with the popular optimizationtechnique called Particle swarm optimization (PSO). The elephant swarm optimizationmodel and the PSO protocol was developed on the SENSORIA Wireless Sensor NetworkSimulator [16] [17] [18] [19] . The elephant swarm optimization algorithm was developeda ݅7 Quad Core CPU having 8GB of RAM to conduct the experimental study.using the C# language on the Visual Studio 2010 platform. The simulations were executed on The experimental set up of wireless sensor network test bed was considered to bespread over a terrain having dimension of 25ൈ25 meters. The incorporating wireless sensornodes were deployed over the environment are varied from 450,500,550,600,650 and 700nodes respectively. The test bed considered sensor nodes mounted with temperature sensorshaving a sensing range of 3m. The communication radio range to be considered is 5m.Sensing events are induced every 0.1 seconds. The induction of such high sensing commotionand deployment of dense networks enables high traffic injection into the test bed. The highertraffic injection that is considered in the test beds results into greater data transactions andultimately resulting into swift energy depletion in the overall considered network. Thesimulation study was conducted for observing the network life time of the test bed. Thethreshold of the network lifetime analysis was set to 30% i.e. the simulation study wasconducted until 30% of the network energy depletes. 55
    • International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 2, March – April (2013), © IAEME In order to justify the efficiency of the elephant swarm optimization algorithm networkthe experimental test beds with varying nodes densities (i.e. 450,500,550,600,650 and 700) wassimulated under predefined and standard specifications provided. The overall depletion innetwork energy was initiated by inducing sensing events. Similarly network topologies wereconsidered to simulate the Elephant Swarm Optimization algorithm and the popular PSOAlgorithm. The simulation period was recorded when the total network energy depleted by 30%.The results obtained are presented in Figure 1 given below. The simulation results achieveddepicts that the cross layer optimization based on the elephant swarm model exhibits a 72.58%improvement in network lifetime when compared to the particle swarm optimization(PSO) basedoptimization protocol. Figure 2 represents the energy decay rate of sensor node with respect to time factor. Herein this depicting resulting figure the results obtained for decay rate has been presented. Theproposed Elephant Swarm optimization model adopts a cross layer approach when compared tothe particle swarm optimization technique. The optimization technique proposed enables thebalanced data scheduling on the links‫ ܮ‬that exist and thus the energy decay rate reduction for thesensor nodes are found to be about 78.67%. NETWORK LIFETIME ANALYSIS ELEPHANT SWARM PSO 500 400 LIFETIME (S) 300 200 100 0 450 500 550 600 650 700 NUMBER OF SENSOR NODES Figure 1: Network Lifetime Analysis WIRELESS SENSOR NODE ENERGY DECAY RATE ELEPHANT SWARM PSO 60 DECAY RATE 40 20 0 450 500 550 600 650 700 NUMBER OF SENSOR NODES Figure 2: Wireless Sensor Node Energy Decay Rate 56
    • International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 2, March – April (2013), © IAEME The proposed elephant swarm optimization technique also enables the communicationoverhead to get reduced even for dense wireless sensor networks. The overall communicationoverheads are reduced greatly by the reduction of the number of retransmissions of datapackets and optimized routing incorporated in the elephant swarm model. Thecommunication overhead of the PSO technique is about 36.6% greater than that of theproposed optimization model. The presented resulting graph depicts that the communicationoverhead is getting increased as the number of sensor nodes are getting increased. The resultsobtained are as shown in Figure 3 of this paper. COMMUNICATION OVERHEAD ELEPHANT SWARM PSO 0.3 COMMUNICATION OVERHEAD 0.25 0.2 0.15 0.1 0.05 0 450 500 550 600 650 700 NUMBER OF SENSOR NODES Figure 3: Communication Overhead Analysis In order to justify the increase in the network lifetime the ratio of the sensor nodesactive at regular time intervals is observed and the results obtained have been presented inTable 1. These results have been obtained for varying sensor node deployment densities. Thegraphical analysis is presented in Figure 4 of this paper given below. The results described inthe table prove that the percentage of active nodes using the elephant swarm optimization isgreater than the nodes alive while using the PSO optimization protocol. Table 1: Active Node Ratio with respect to the Simulation Instance and Network Topology SizeNUMBER OF SIM TIME ELEPHANT SWARM PSO ACTIVE NODE SENSOR (S) ACTIVE NODE RATIO RATIO NODES 450 293 99.77777778 79.55555556 500 335 99.8 80.6 550 240 99.81818182 77.27272727 600 193 99.83333333 80.66666667 650 222 99.84615385 77.53846154 700 209 99.85714286 76.42857143 57
    • International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 2, March – April (2013), © IAEME The experimental study discussed in this section of this paper proves that the elephantswarm optimization technique proposed in this paper exhibits better network performancethan the popular Particle Swarm Optimization (PSO) protocol commonly used. The enhancednetworks performance is proved in terms of network lifetime, communication overheadreduction, reduced energy decay in the sensor nodes and enhanced active node ration in thenetwork. % OF ACTIVE NODES 120 ELEPHANT SWARM PSO 100 ACTIVE NODE RATIO% 80 60 40 20 0 450 500 550 600 650 700 NETWORK SIZE Figure 4: Active Node Ratio in the Test Bed Observed At Constant Simulation Slots5. CONCLUSION AND FUTURE WORK In this manuscript the authors address the problem in enhancing the network lifetimeof wireless sensor networks. The elephant swarm optimization technique is adopted toaddress the issue that exists. A cross layer approach is adopted to incorporate optimizations atthe routing, radio and the MAC layers. A TDMA based MAC layer is considered and theMAC schedule is optimized in accordance to the routing and the radio link layeroptimization. The system model considered is clearly discussed. The optimization functionwhich needs to be solved using the elephant model is also discussed. The elephant swarmoptimization is achieved using a phased approach discussed in this paper. The experimentalevaluation conducted proves the efficiency of the proposed elephant swarm optimizationtechnique over the optimization technique called Particle swarm optimization (PSO)technique in terms of improved network lifetime, reduced sensor node energy decay rate,higher active node ratios and lower communication overheads. The overall network lifetimeof the varied scenarios presented proves enhancement of about 26.8% thus justifying therobustness of the proposed elephant swarm optimization technique. The future of this work can be considered to compare the elephant swarmoptimization technique with other evolutionary computing based optimization techniques likeModified Interactive based Evolutionary Computing (MIEC) techniques with real timedeployment parameters and then justify its robustness over the existing systems. 58
    • International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 2, March – April (2013), © IAEMEREFERENCES[1] Wilson, E.O. 2000. Sociobiology - The New Synthesis. Twenty Fifth Anniversary Edition.The Belknap Press of Harvard University Press Cambridge, Massachusetts and London,England.[2] Krebs, J. R., Davies, N. B. 1993. An Introduction to Behavioural Ecology - ThirdEdition. Blackwell Publishing. Oxford, UK.[3] Elizabeth A. Archie, Cynthia J. Moss and Susan C. Alberts, "The ties that bind: geneticrelatedness predicts the fission and fusion of social groups in wild African elephants.",Proc.R. Soc. B (2006) 273, 513–522[4] ELIZABETH A. ARCHIE and PATRICK I. CHIYO, "Elephant behaviour andconservation: social relationships, the effects of poaching, and genetic tools formanagement", Molecular Ecology (2012) 21, 765–778[5] Rachael Adams ,"Social Behaviour and Communication in Elephants- Its true! Elephantsdont forget! " Available at : http://www.wildlife-pictures-online.com/elephant-communication.html Acessed March 14th 2013[6] Wyatt, T.D. 2003. Pheromones and Animal behaviour - Communication by Smell andTaste. Cambridge University Press. UK[7] Pacharintanakul, Peera; Tipper, David W., “Differentiated cross layer network mapping inmultilayered network architectures”, Telecommunications Network Strategy and PlanningSymposium (NETWORKS), 2010 14th International on 27-30 Sept.[8] Kamatam, G.R.; Srinivas, P.V.S.; Chandra Sekaraiah, K., “SC3ERP: Stratified CrossLayer Congestion Control and Endurance Routing Protocol for Manets”, ComputationalIntelligence, Communication Systems and Networks (CICSyN), 2012 Fourth[9] Jin, Lizhong; Jia, Jie; Chen, Dong; Li, Fengyun; Dong, Zhicao; Feng, Xue; “Research onArchitecture, Cross-Layer MAC Protocol for Wireless Sensor Networks”, Genetic andEvolutionary Computing (ICGEC), 2011 Fifth International Conferenc[10] Zheng Guoqiang; Ning Yonghai; Tang Shengyu; “Cross layer-based MAC protocolfor wireless sensor networks”, Control Conference (CCC), 2010 29th Chinese, 29-31 July2010, Page(s): 4748 – 4752.[11] Berry, Randall A.; Yeh, Edmund M.; “Cross-layer wireless resource allocation”,Signal Processing Magazine, IEEE on Sept. 2004, Volume: 21 , Issue:5 Page(s): 59 – 68.[12] Hang Su; Xi Zhang; “Cross-Layer Based Opportunistic MAC Protocols for QoSProvisioning Over Cognitive Radio Wireless Networks”, IEEE Journal on Selected Areas inCommunication on Volume 26 Issue 1, January 2008, Page 118-129.[13] I.F. Akyildiz; W. Su, Y; Sankara subramaniam; E.Cayirci, “Wireless sensor networks:a survey”, Computer Networks 38 (2002) 393–422.[14] Supiya Ujjin;Peter J. Bentley; “Particle Swarm Optimization Recommender System”,2003.[15] G. J. Foschini and J. Salz, “Digital communications over fading radio channels,” BellSystems Technical J., pp. 429–456, Feb. 1983.[16] A. K. Dwivedi and O. P. Vyas, "An Exploratory Study of Experimental Tools forWireless Sensor Networks",Wireless Sensor Network, 2011, 3,pp 215-240.[17] Laiali Almazaydeh, Eman Abdelfattah, Manal Al- Bzoor, and Amer Al- Rahayfeh,"Performance Evaluation Of Routing Protocols In Wireless Sensor Networks", InternationalJournal of Computer Science and Information Technology, Volume 2, Number 2, April 2010pp 64-73. 59
    • International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print), ISSN 0976 – 6375(Online) Volume 4, Issue 2, March – April (2013), © IAEME[18] Jamal N. Al-Karaki and Ghada A. Al-Mashaqbeh, "SENSORIA: A New SimulationPlatform for Wireless Sensor Networks", 2007 International Conference on SensorTechnologies and Applications, 2007 IEEE DOI 10.1109/SENSORCOMM.2007.43.[19] Amer A. Al-Rahayfeh, Muder M. Almi’ani, and Abdelshakour A.Abuzneid,"Parameterized Affect Of Transmission Range On Lost Of Network Connectivity(Lnc) Of Wireless Sensor Networks",International Journal of Wireless & Mobile Networks (IJWMN ), Vol.2, No.3, August 2010. pp 63-80.[20] Bharathi M A, Vijaya Kumar B P and Manjaiah D.H, “Power Efficient DataAggregation Based on Swarm Intelligence and Game Theoretic Approach in Wireless SensorNetwork” International journal of Computer Engineering & Technology (IJCET), Volume 3,Issue 3, 2012, pp. 184 - 199, ISSN Print: 0976 – 6367, ISSN Online: 0976 – 6375, Publishedby IAEME.[21] Basavaraj S. Mathapati, Siddarama. R. Patil and V. D. Mytri, “Power Control WithEnergy Efficient and Reliable Routing MAC Protocol for Wireless Sensor Networks”International journal of Computer Engineering & Technology (IJCET), Volume 3, Issue 1,2012, pp. 223 - 231, ISSN Print: 0976 – 6367, ISSN Online: 0976 – 6375, Published byIAEME[22] Poonam Thakur and M.Vijaya Raju, “Survey on Routing Techniques for Manets andWireless Sensor Networks: A Comparison” International journal of Computer Engineering &Technology (IJCET), Volume 4, Issue 1, 2013, pp. 275 - 283, ISSN Print: 0976 – 6367,ISSN Online: 0976 – 6375, Published by IAEME. 60