Mobile Ad-hoc Networks (MANETs) are multi-hop ad hoc networks of mobile battery-charged devices designed for emergency use. There is source preparation with regards to a potential disconnection whilst occurrence of a path breakage. The source will commence with discovery of the path and evade complete disconnection. When there is a decrease in the availability of links, there may be path breakage. Breakages of links because of energy drain and node mobility are managed by reactive MANET routing protocols. Also, multi-hop communication can be more efficient in terms of the routing protocols for large area networks. The Ad hoc On-Demand Multipath Distance Vector protocol (AOMDV) is modified by this study in a way that link availability estimates lead to route handoff. An optimal AOMDV route can be detected by the novel Bat Algorithm hybridization with Harmony Search (BAT-HS) algorithm depending on rate of bit error, velocity, neighbouring node’s queuing delay, and link quality. In this work, the proposed algorithm Hybrid Bat (HBAT) to enhanced the AOMDV with Hybrid Bat is optimizes the routing outperforming the Packet Delivery Ratio (PDR), End to End delay reduction and in order to reduce routing control overheads and route discovery that are proved by simulation results.
Research Article
An Enhanced Ad hoc On-Demand Multipath Distance Vector Routing using Hybrid Bat Algorithm
https://doi.org/10.21203/rs.3.rs-1097752/v1
This work is licensed under a CC BY 4.0 License
Version 1
posted
You are reading this latest preprint version
AOMDV
Link Availability
Hybrid BAT Algorithm
Harmony Search Algorithm
Packet Delivery Ratio
Route discovery
Quality of Service (QoS)
MANETs are wireless networks with self-configuration by mobile routers that are wireless link-connected to set up an absolute topology. Routers are the participating nodes which have free and random motion. These nodes also arbitrarily manage themselves in the router. In a wireless network topology, this leads to swift variations. These networks may also be connected to a bigger Internet. Quality of Service (QoS) provisioning’s purpose is achievement of deterministic network behaviour for better utilization of network resources as well as better information delivery by the network [1].
Host and router in MANETs are the mobile nodes for packet forwarding to other nodes since they may not be within each other’s ranges of transmission [1]. Reactive and proactive routing protocols are classifications of MANET routing which are route discovery time dependent. There is systematic route discovery and renewal in table-driven (proactive) routing protocols. Node routing tables are employed for storage of every data related to routing. A traffic source utilizes the routing table’s accessible route when required. However, in on-demand routing protocols, the process of route discovery is admitted by the traffic source only when there is a route requirement [2].
The reactive or on-demand route discovery is offered by AODV. Route discovery will identify paths from source to destination and also when routing information is not present in the source node’s routing table. The neighbours are broadcasted with a route request (RREQ) message by a node for requesting a destination route. The RREQ packet directly broadcasts a path replay (RREP), which is a unicast reply message, to the source from which the RREQ message was received [3].
A multi-path AODV extension is the AOMDV routing protocol. Three aspects of the AOMDV are comparable with the on-demand multi-path routing protocols. At first, it may not seem to have intermodal coordination overheads like other protocols. Later, without source routing, it can secure alternative route disjointness via distributed calculation. In the end, AOMDV will assess substitute routes with nominal extra overhead over AODV through potential alternate path routing data utilization [2].
Links with high reliability for participating nodes in route establishment/maintenance are ensured by the routing scheme. Ergo, MANETs face the issue of designing reliable routing protocols. Physical conditions, mobility, received signal strength, and node power are factors that depend on link reliability. Surveillance applications, audio/video conferencing are a few groups of communication services that require reliable nodes along the multicast path to continuously transmit non-prioritized/prioritized packets. Redundancy in sensing/communicating nodes help in the accomplishment of reliability in the nodes of these applications [4].
For research purposes, routing appears as a prominent issue. An optimal path between a source/destination set is decided by the frequently used single path routing protocols. It is required to find new paths during route breaks which results in high latency/overhead [7]. During route maintenance, breakages in the MANET links are identified by AOMDV and AODV. After this, there is election of a new route. There is delay, active path’s packet loss, and performance degradation because of the time taken for a new route’s election. For instances of breakages because of lower energy level and mobility, there is analysis of AODV/AOMDV. Whilst the AODV places multiple paths on route, the AOMDV places multiple paths in the routing table.
Alternate route election on link failure must not cause a new route discovery initiation. Without loss of data, there is election of an alternate path from the succeeding hop list [5]. The AOMDV routing protocol may be potentially optimized by maintenance of alternate routes which are found at the time of route discovery, and efficient load distribution so as to prevent overburdening of the nodes. This holds true only during an alternate route’s availability. Transmission of data along alternate routes will maintain them as well as prevent timeout issues [6].
In this work Bat Algorithm is applied for AOMDV’s enhancement for the occurrence of route handoff utilizing the estimation of link availability. Bat Algorithm also has other variations and enhancements. A hybrid Bat Algorithm for function benchmarks’ numerical optimization is a duct of the Hybridized Bat Algorithm with Harmony Search [8]. Depending on factors like rate of bit error, velocity, neighbouring node’s queuing delay, and link quality, the novel hybrid optimization algorithm based on Bat Algorithm and Harmony Search (BAT-HS) will identify an optimal AOMDV route. The article is divided into different sections; Section 2 discusses AOMDV optimization works, Section 3 describes the various techniques on which the proposed method is based, Section 4 presents the results and conclusion in Section 5.
Routing process control and leaving of hops which only had maximum permissible distances in a route that relied on each node’s received signal strength was proposed by Lalitha & Rajesh [9]. A maximum transmission range-based, power-ware, and topology-aware AOMDV routing protocol was recommended by the authors. This AOMDV_Range Routing (RR) model was NS2 deployed by the standard AOMDV protocol’s improvement. Testing was done on the standard AOMDV and the latest AOMDV_RR under diverse network densities, and there was inclusion of the performances for metrics that were favourable. There was realization of measurable performance differences. Compared to the standard AOMDV, the AOMDV_RR demonstrated improved QoS like higher throughput, and lower network overhead and also conservation of energy.
Avoidance of congestion by better route identification could boost the AOMDV’s performance by combining with Sufficient Bandwidth Aware (AOMDV+SBA) which was proposed by Wannawilai & Sathitwiriyawong [10]. Outcomes of the NS2 simulations demonstrated increased PDR below medium-to-high traffic loads, decreased power consumption, decreased data packet delay, and network performance enhancement and endurance. The proposed protocol’s performance in low-to-medium traffic loads, was near to that of the standard AOMDV routing. NS2 Simulations were used for analysis of differences in the performance of the AOMDV and the Optimized Link State Routing (OLSR) routing protocol by Oo & Othman [11]. The authors compared the benefits and drawbacks of both protocols using normalised routing load, and measured packet loss.
Modified RREP in order to select dependable routes based on Bit Error Rate (BER) was done by Yelemou [12]. There was study of the effectiveness of this protocol with regards to terms of these developments under realistic conditions. Comparisons were also drawn against the standard AOMDV/AODV protocols. Outcomes demonstrated that, even in difficult conditions such as multi-communication or mobility, they could boost the standard AOMDV’s performance.
There was a presentation of a scheme that was stable link-based and consisting of a mobility prediction-based node election process and a battery power ratio by Kant & Awasthi [13]. This scheme could set up stable routes for carrying out routing discovery, broadcasting messages, packet transmission, and maintenance of multicast routing in MANETs. For illustration of the consequence of node failure on MANET performance, Choudhary et al., [14] has proposed a node failure model incorporated in the network. This was represented as a node failure model to assess the behaviour of the MANETs in scenarios like variations in topology, and realistic momentary node failures. Gap analysis execution and strategy preparations were done to boost the reliability and the performance of MANET.
A novel AOMDV protocol that contains an energy threshold for optimization of the forwarding technique whilst simultaneously constructing an energy-cost function for utilization as a multiple path selection criterion. For balancing the multiple path energy at the time of transmission, it employs multiple paths one by one. NS2 simulation tool compares the performances of AOMDV and MRNLM Liu et al., [15].
The misbehaviour in Transmission Control Protocol (TCP) was examined in detail by Kathiravan & Radha [16]. Based on their analysis, a cross layer mechanism which employed AOMDV was proposed by them. In the protocol, multiple link disjoint paths were identified utilizing AOMDV and the next-hop was chosen based on existing wireless channel conditions for wireless channel error minimization. When there was failure in the initial route failure, new window size for a substitute route was preferred by the TCP source. The distance between the alternate path/destination’s origins formed the basis of this opportunistic window size election. NS2 was used to evaluate this protocol’s performance with the AOMDV for metrics like BER, PDR, size of window, and TCP throughput. Significant developments were observed by assessing these outcomes.
Performance optimization of the AODV routing protocol was done with usage of link availability prediction in urban areas via Ghanbarzadeh & Meybodi’s [17] Hello message mechanism for modelling urban areas with the Manhattan mobility model for movement of nodes. Presentation on link availability probability was done by Kumar et al., [18]. For simulation work, the Nakagami model is employed, and the model is evaluated in an urban environment with a variety of velocity between nodes. The link availability between nodes is also determined.
Yuet al., [19] presented Link Effective Available Time (LEAT) protocol for getting available link-time through an epoch through estimation of distance between nodes instead of utilizing the total localization information. Then, for link breakage reduction, the product of link availability and available time was proposed as a novel link cost for routing. Simulations confirmed that LEAT could support network performance in throughput, hop count, and delay, which are associated with well-known routing algorithms. LEAT could also significantly minimize the link breakages
Investigations were done by Houaidia et al., [20] on effective routing metrics design for optimizing network resources and for fulfilling the QoS requirements of users. Actual experiments were conducted to uncover the Expected Transmission Count (ETX) metric’s deficiency in subsequent optimization for routing that was effective. The OLSR routing protocol was used to perform simulations. These simulations demonstrated the ETX’s various deficiencies that led to inaccurate link quality estimation and routing decisions. Link availability-based enhancements for the ETX metric were presented in this work for accurate detection of high-throughput paths in wireless mesh networks which are multi-hop.
In this work, the AOMDV is optimized using a heuristic method to find the optimal route. The proposed technique is based on Hybrid BAT optimization. To find the optimal route, the link quality, BER, node velocity and neighbouring node queuing delay are considered. In this section methods like link availability and Hybrid BAT Harmony Search Algorithm are presented.
3.1 Link Availability Estimation
If nodes A and B have an active link at time t, link availability being active between them at time t+T. A link can recover or fail one or more times in the time span [t, t+T]. For path selection between nodes, this metric is not a practical criterion. This is because, if there is link failure, there must be immediate rerouting instead of waiting for a failed link’s re-availability. This evaluation’s concept is summarized below [21]:
Provided that T is the prediction of expiration time (time available continuously) for an active link {a, b} between A and B at time t. Then, the definition for availability L(T) of link {a, b} will be:
L(T) = PR{link{a, b} active between time t to t+T given that it was active at time t}
-
A prediction-based link viability prediction allows a node to provide a continual time period that an accessible link lasts from beginning time, assuming that both link nodes have no alterations and maintain their current motion in speed and distance. Below are the basic assumptions in the derivation of continuous link availability: [25].
-
The definition of an epoch is given as the random length interval when a node traverses with constant velocity.
-
For the sake of simple discussions, the assumption is that a node will have the same mean epoch length \({\lambda ^{ - 1}}\). Despite that, the derivation, \(L\left( {{t_p}} \right)=P\left( {{{T>\left( {{t_0}+{t_p}} \right)} \mathord{\left/ {\vphantom {{T>\left( {{t_0}+{t_p}} \right)} {T>{t_0}}}} \right. \kern-0pt} {T>{t_0}}}} \right)\)can be extended to different mean epoch lengths for nodes in which \({t_0}\) is the initial time from when the prediction begins.
-
The prediction \({t_p}\) is given on continuous availability of a link\(L\left( {{t_p}} \right)\) between two exponential nodes which are identical.
-
Uncorrelation of node mobility [22].
3.2 Enhanced AOMDV Routing with Hybrid Bat
Sonar echoes are utilized by bats for obstacle detection. Short, loud sound pulses (with pulse rates of 0 to 20 times/second) are emitted by the bats during their flight. Upon hitting the obstacles, the sound pulses undergo transformation into a frequency. These are then reflected upon hitting any object. The bats use the delay between the emission and reflection for navigation. Bats will convert their pulses into useful information to assess the distance from their prey. Wavelengths with a range of 0.7 to 17 mm or inbound frequencies of 20-500 kHz are utilized by the bats [26].
Harmony Search (HS) is developed on the basis of natural musical performance processes that take place when an optimal harmony state is sought by the musician. Operators that are used for specifying the HS algorithm optimization are: bandwidth of pitch adjustment (bw), rate of pitch adjustment (PAR), rate of harmony memory consideration (HMCR), Harmony Memory (HM) and the harmony memory size (HMS) - it shows the solution vector stored in the HM.
HS was motivated by the improvisation of Jazz musicians who individually refine their individual improvisation through a piece of music variation resulting in musical variations that produce aesthetic harmony. Stages of HS protocol are:
Begin
HM initialized and fitness evaluated.
do
for 1-D in d do
Compare rand and HMCR // memory
assign 𝑥𝑎 to 𝑥new of (𝑑)
Compare rand and PAR // pitch adjustment
𝑥new (𝑑) = 𝑥old (𝑑) + 𝑏𝑤 × (2 × rand − 1)
end
𝑥new(𝑑) = 𝑥min,𝑑+ rand × (𝑥max,𝑑− 𝑥min −𝑑)
end
end for d
Updating HM as 𝑥𝑤 = 𝑥new, if (𝑥new) < (𝑥𝑤) (minimization objective)
Updating best harmony vector
Step 4. end
End
The steps of the protocol are detailed below:
Stage 1. Setting up of the optimization issue as well as algorithm variables
In the initial stage, the optimization issue is given by
Min (or Max ) \(f(\underset{\_}{x)}\)subject to\({x}_{i}ϵ{X}_{i}, i =\text{1,2},....,N.\)
The solution is obtained by either maximizing or minimizing as required. In this work, for a specified source to destination, several paths are identified through the link measure in AOMDV by modifying path discovery procedure. For maintaining a balance between network loads as well as QoS, the aim is the minimization of packet loss rate, estimated load, and the delay in route.
F(i) are scalar objective functions to be optimized Xi refers to the range of values of all continuous decision variables xi.
Additionally, the control variables of HS are also given in this particular stage. The variables are the HMS that is the quantity of solution vectors (Chakraborty et al., 2009 and Amiri et al., 2010).
Stage 2. HM initialization
Every member of every matrix in the concurrent population of HM that is the extent of HMS is assigned a random integer that is equitably spread between upper and lower bounds at this phase.
The following equation is used to achieve this for the i-th element of the j-th solution vector:
$$[{L}^{{x}_{i}},{U}^{{x}_{i}}], 1\le i\le N$$
$${x}_{i}^{j}={L}^{{x}_{i}}+rand\left(\text{0,1}\right).({U}^{{x}_{i}}-{L}^{{x}_{i}})$$
Stage 3. New Harmony improvisation:
A Harmony vector \(\underset{\_}{x }=({x}_{1}^{\text{'}},{x}_{2}^{\text{'}},....{x}_{N}^{\text{'}})\) is created on the based on
(1) Memory,
(2) Pitch adjustments, and
(3) Arbitrary selection.
Creating a novel harmony known as ‘improvisation’ helps to explore the solution space for better solutions.
Stage 5. Check terminating criteria: Last stage is to check the criterion for continuing the process the terminating criteria (maximal NI) is fulfilled, computations are stopped. Else, stages 3 and 4 are iterated.
Although the Bat Algorithm is effective in exploitation (local search), it does fall into the local optima sometimes. As a result, it loses its capability to carry out proper global search. As BAT relies on random walks for searching, it is unable to ensure rapid convergence. By hybridizing BAT with HS by maximizing BAT’s population diversity, the issue of falling into the local optima can be avoided. Addition of pitch adjustment in HS is an enhancement which acts as a mutation operator for making convergence faster so as to extend the approach’s feasibility for even more practical applications whilst ensuring that the basic BAT’s attractive characteristics are preserved.
A hybrid metaheuristic approach that incorporates pitch adjustment in HS as an evolutionary algorithm into the BAT allows for benchmark function optimization. The distinction between HS/BAT and BAT is that the mutation operator improves the original BAT, resulting in a unique solution for each bat [26]. As a result, this strategy will use HS algorithms to seek out new search space, as well as BAT to leverage population information in order to avoid falling into the local optima in BAT [25, 27].
The BAT-HS critical operator is a hybrid harmony search mutation operator that improvises harmony in HS with BA. Following considerations form the basis behind this new hybrid mutation operator: the new search space’s exploration is enhanced by the mutation operator. Thus, there is full development of the BAT’s exploitation capabilities as well as the strong exploration capabilities of the original HS.
Bat Algorithm (BAT) is unable to ensure rapid convergence as the search wholly depends on random walks. BAT can be enhanced by the addition of the mutation operator which also includes three minor enhancements for speeding up convergence so as to make the method feasible for practical applications whilst retaining the attractive feature of the original method.
The simulations were performed on the MATLAB platform. the number of nodes 200 to 1000 nodes have been experimented with and the transmit power of each node is 0.005 watts. Table 1. shows the simulation parameters. The network is simulated and the proposed algorithm is evaluated for the achieved packet delivery speed and the end-to-end delay. Table 2 to 4 and Figure 3 to 8 illustrate the results obtained.
| Parameter | Values |
|---|---|
| Number of nodes | 200-1000 |
| Simulations area (m) | 1000 x 1000 |
| Transmission range (m) | 100 |
| Traffic model | CBR |
| Packets size (bytes) | 512 |
| Simulation time | 2000 sec |
| Bandwidth | 2Mbps |
| Node pause(Sec) | AOMDV | LQ AOMDV | BAT AOMDV | HBAT AOMDV |
|---|---|---|---|---|
| 200 | 0.7391 | 0.7333 | 0.7381 | 0.7542 |
| 400 | 0.7597 | 0.8617 | 0.8871 | 0.9171 |
| 600 | 0.7852 | 0.8741 | 0.8889 | 0.9043 |
| 800 | 0.8442 | 0.9509 | 0.9657 | 0.9712 |
| 1000 | 0.9191 | 0.9643 | 0.978 | 0.9956 |
Table 1 and figure 2 it is observed that for node pause time of 200, Packet Delivery Ratio of HBAT-AOMDV improves by 5.49% than LQ-AOMDV and by 3.54% than BAT-AOMDV. For node pause time of 800, Packet Delivery Ratio of HBAT-AOMDV improves by 2.11% than LQ-AOMDV and by 0.57% than BAT-AOMDV. For node pause time of 1000, Packet Delivery Ratio of HBAT-AOMDV improves by 1.79% than LQ-AOMDV and by 0.38% than BAT-AOMDV.
| Node pause (Sec) | AOMDV | LQ AOMDV | BAT AOMDV | HBAT AOMDV |
|---|---|---|---|---|
| 200 | 0.003833 | 0.004191 | 0.004465 | 0.004085 |
| 400 | 0.00218 | 0.001128 | 0.002497 | 0.002367 |
| 600 | 0.002012 | 0.001624 | 0.001259 | 0.001216 |
| 800 | 0.000812 | 0.001221 | 0.00057 | 0.000547 |
| 1000 | 0.001087 | 0.000785 | 0.000754 | 0.000702 |
Table 2 and figure 3 it is observed that for node pause time of 200, Jitter of HBAT-AOMDV is reduced by 0.15% than LQ-AOMDV and by 6.63% than BAT-AOMDV. For node pause time of 800, Jitter of HBAT-AOMDV is reduced by 76.24% than LQ-AOMDV and by 4.12% than BAT-AOMDV. For node pause time of 1000, Jitter of HBAT-AOMDV is reduced by 11.16% than LQ-AOMDV and by 7.14% than BAT-AOMDV.
| Node pause (Sec) | AOMDV | LQ AOMDV | BAT AOMDV | HBAT AOMDV |
|---|---|---|---|---|
| 200 | 0.8709 | 0.9235 | 0.8262 | 0.7706 |
| 400 | 0.5107 | 0.4711 | 0.485 | 0.4687 |
| 600 | 0.4035 | 0.3318 | 0.377 | 0.3598 |
| 800 | 0.2815 | 0.3076 | 0.2606 | 0.246 |
| 1000 | 0.207 | 0.265 | 0.1928 | 0.1836 |
Table 3 and Figure 4 it is observed that for a node pause time of 200, the mean end-to-end delay of HBAT-AOMDV is reduced by 18.16% than LQ-AOMDV and by 7.04% than BAT-AOMDV. For node pause time of 800 the end-to-end delay of HBAT-AOMDV is reduced by 22.25% than LQ-AOMDV and by 5.46% than BAT-AOMDV. For node pause time of 1000 the average end-to-end delay of HBAT-AOMDV is reduced by 36.29% than LQ-AOMDV and by 4.89% than BAT-AOMDV.
In Figure 5 it is observed that the routing control overhead of proposed HBAT-AOMDV is decreased when contrasted with BAT-AOMDV and LQ-AOMDV and incremented when contrasted with AOMDV. Average values of BAT-AOMDV are increased by 18.8% than AOMDV but decreased by 11.76% than LQ-AOMDV. Average values of HBAT-AOMDV are increased by 20.4% than AOMDV but decreased by 13.52% than LQ- AOMDV.
Figure 6 shows that the HBAT-AOMDV number of route discoveries has decreased on average from 7.65–32.09% when contrasted with AOMDV, LQ AOMDV and BAT AOMDV.
In Figure 7 it is seen that the HBAT-AOMDV for routing/Data bits decreases by the average of 0 to 26.08% as opposed to AOMDV, LQ-AOMDV and BAT-AOMDV.
To avoid node overload, there is maintenance of alternative routes found during path discovery. and effective load distribution in the AOMDV routing protocol. Positive outcomes were demonstrated when performance comparisons were done against commonly known population-based algorithms. Simulations are conducted to evaluate the Hybrid Bat AOMDV (HBAT-AOMDV) and compared with LQ AOMDV and BAT AOMDV. The experiment was performed with 1000 nodes and the transmitting power of each node was 0.005 watts. Simulation results of this work for PDR, end to end delay and jitter achieved. The proposed HBAT AOMDV optimizes the routing based on the link availability and is based on hybridization of BAT and HS algorithms. In comparison to other algorithms, high-quality solutions are found by the Hybrid BAT-HS. The results show the hybrid algorithm compared to traditional, the average end to end delay of Hybrid BAT AOMDV is improved by reducing delay as opposed to LQ AOMDV and BAT AOMDV. Average values of HBAT AOMDV are reduced by 12.54% than LQ AOMDV and reduced by 5.4% than BAT AOMDV. It is observed that the average jitter of Hybrid BAT AOMDV reduces jitter as opposed to LQ AOMDV and BAT AOMDV. Average jitter for HBAT AOMDV is reduced by 0.77% than LQ AOMDV and reduced by 5.75% than BAT AOMDV. Future work could include integrating the proposed solutions into a single framework and developing new techniques based on fuzzy mathematical programming for multi-metric optimization, taking variations into account.
Conflict of Interests
The authors proclaim that there is no conflict of interests concerning the publication of this paper.
*No, Funding (information that explains whether and by whom the research was supported).
*No, Availability of data and material (data transparency) .
*No, Code availability (software application or custom code).
*Yes, Authors' contributions (optional: please review the submission guidelines from the journal whether statements are mandatory).
Dr. R.Prabha : Conceptualization , Methodology , Visualization, Supervision, Investigation, Algorithm & Flow diagram.
Senthil G A: Related Work, Software, Data Validation, Writing- Original Draft preparation, Writing -Reviewing and Editing.
- Gupta, D., & Gujral, R. K. Simulation of Different Routing Protocols in MANET Using NS2
- Kute, V. B., & Kharat, M. U. (2012). Analysis of Quality of Service for the AOMDV Routing Protocol. Engineering, Technology & Applied Science Research, 3(1), pp-359
- Saadi, Y., Kafhali, S. E., Haqiq, A., & Nassereddine, B. (2013). Simulation analysis of routing protocols using manhattan grid mobility model in MANET.arXiv preprint arXiv:1304.2035
- Biradar, R. C., & Manvi, S. S. (2011). Information Priority Based Multicast Routing in MANETs. International Journal of Wireless & Mobile Networks (IJWMN) Vol, 3, 177–192
- Murali, L., & Divya, T. V. (1956/V2N1-148). Impact of link breaks on AODV and AOMDV in MANET: A Performance Analysis. International Journal of Electronics and Computer Science Engineering (IJECSE), ISSN 2277- 153
- Das, S., Roy, R., & Das, P. K. (2006). Optimizations To Multipath Routing Protocols In Mobile Ad hoc Networks. In Proc International Conference on Emerging Applications of IT (EAIT 2006), Kalkata, India
- Kalpana, R., & Baskaran, M. (2016). TAR: TOA–AOA based random transmission directed localization. Wireless Personal Communications, 90(2), 889–902
- Yang, X. S., & He, X. (2013). Bat algorithm: literature review and applications. International Journal of Bio-Inspired Computation, 5(3), 141–149
- Lalitha, V., & Rajesh, R. S. (2014, January). Power aware and topology aware ad-hoc on-demand multipath distance vector routing for MANET. In Intelligent Systems and Control (ISCO), 2014 IEEE 8th International Conference on (pp. 115-119). IEEE
- Wannawilai, P., & Sathitwiriyawong, C. (2010, June). AOMDV with sufficient bandwidth awareness. In Computer and Information Technology (CIT), 2010 IEEE 10th International Conference on (pp. 305-312). IEEE
- Yelemou, T. (2013, March). New approach to improve the robustness of AOMDV protocol. In Systems, Signals & Devices (SSD), 2013 10th International Multi-Conference on (pp. 1-6). IEEE
- Kant, K., & Awasthi, L. K. (2010, November). Stable link based multicast routing scheme for MANET. In Computational Intelligence and Communication Networks (CICN), 2010 International Conference on (pp. 296-300). IEEE
- Choudhary, A., Roy, O. P., & Tuithung, T. (2014, April). Node failure effect on reliability of mobile ad-hoc networks. In Communication Systems and Network Technologies (CSNT), 2014 Fourth International Conference on (pp. 207-211). IEEE
- Liu, J., Chen, J., & Kuo, Y. (2009, September). Multipath routing protocol for networks lifetime maximization in ad-hoc networks. In Wireless Communications, Networking and Mobile Computing, 2009. WiCom'09. 5th International Conference on (pp. 1-4). IEEE
- Kathiravan, K., & Radha, R. (2012, April). An opportunistic cross layer multi path routing protocol to improve TCP performance over Mobile Ad Hoc Networks. In Recent Trends In Information Technology (ICRTIT), 2012 International Conference on (pp. 320-325). IEEE
- Ghanbarzadeh, R., & &Meybodi, M. R. (2010, May). Reducing message overhead of AODV routing protocol in urban areas by using link availability prediction. In Computer Research and Development, 2010 Second International Conference on (pp. 274-279). IEEE
- Ghanbarzadeh, R., & &Meybodi, M. R. (2010, May). Reducing message overhead of AODV routing protocol in urban areas by using link availability prediction. In Computer Research and Development, 2010 Second International Conference on (pp. 274-279). IEEE
- Kumar, S., Javaid, N., Yousuf, Z., Kumar, H., Khan, Z. A., & Qasim, U. (2012, October). On link availability probability of routing protocols for urban scenarios in VANETs. In Open Systems (ICOS), 2012 IEEE Conference on (pp. 1-6). IEEE
- Houaidia, C., Van Den Bossche, A., Idoudi, H., Val, T., & Saidane, L. A. (2013, May). Link availability aware routing metric for wireless mesh networks. InComputer Systems and Applications (AICCSA), 2013 ACS International Conference on (pp. 1-4). IEEE
- Gavalas, D., Konstantopoulos, C., & Pantziou, G. (2010). Mobility Prediction in Mobile Ad Hoc Networks.Next Generation Mobile Networks and Ubiquitous Computing,226–240
- Akkaya, K., & Younis, M. (2003, May). An energy-aware QoS routing protocol for wireless sensor networks. In Distributed Computing Systems Workshops, 2003. Proceedings. 23rd International Conference on (pp. 710-715). IEEE
- Senthil, G. A., Raaza, A., & Kumar, N. (2021).). Internet of Things Energy Efficient Cluster-Based Routing Using Hybrid Particle Swarm Optimization for Wireless Sensor Network. Wireless Pers Commun( https://doi.org/10.1007/s11277-021-09015-9
- Senthil, G. A., Raaza, A., & Kumar, N. (2021). Internet of things multi hop energy efficient cluster-based routing using particle swarm optimization. Wireless Netw. https://doi.org/10.1007/s11276-021-02801-0.
- Kumar, V. S. (2014). Hybrid optimized list scheduling and trust-based resource selection in cloud computing.Journal of Theoretical & Applied Information Technology, 69(3)
- Prabha, R., & Ramaraj, N. (2015). "An improved multipath MANET routing using link estimation and swarm intelligence",EURASIP Journal on Wireless Communications and Networking,
- Geem, Z. W., Kim, J. H., Loganathan, G. V., & Simulation (2001). vol. 76, no. 2, pp.60–68,
- Lenin, K., Reddy, B. R., & Kalavathi, M. S. (2014). Diminution of Real Power Loss by Using Hybridization of Bat Algorithm with Harmony Search Algorithm. International Journal of Computer (IJC), 15(1), 59–78