The Scientific World Journal

Volume 2015 (2015), Article ID 284276, 9 pages

http://dx.doi.org/10.1155/2015/284276

## Energy-Aware Multipath Routing Scheme Based on Particle Swarm Optimization in Mobile Ad Hoc Networks

^{1}Department of Computer Science and Engineering, SCAD College of Engineering and Technology, Tirunelveli, Tamil Nadu 627414, India^{2}Anna University, Chennai, Tamil Nadu 600 025, India

Received 17 September 2015; Accepted 24 November 2015

Academic Editor: Muthu Ramachandran

Copyright © 2015 Y. Harold Robinson and M. Rajaram. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

#### Abstract

Mobile ad hoc network (MANET) is a collection of autonomous mobile nodes forming an ad hoc network without fixed infrastructure. Dynamic topology property of MANET may degrade the performance of the network. However, multipath selection is a great challenging task to improve the network lifetime. We proposed an energy-aware multipath routing scheme based on particle swarm optimization (EMPSO) that uses continuous time recurrent neural network (CTRNN) to solve optimization problems. CTRNN finds the optimal loop-free paths to solve link disjoint paths in a MANET. The CTRNN is used as an optimum path selection technique that produces a set of optimal paths between source and destination. In CTRNN, particle swarm optimization (PSO) method is primly used for training the RNN. The proposed scheme uses the reliability measures such as transmission cost, energy factor, and the optimal traffic ratio between source and destination to increase routing performance. In this scheme, optimal loop-free paths can be found using PSO to seek better link quality nodes in route discovery phase. PSO optimizes a problem by iteratively trying to get a better solution with regard to a measure of quality. The proposed scheme discovers multiple loop-free paths by using PSO technique.

#### 1. Introduction

A MANET [1] is composed of mobile nodes connected by wireless media without centralized infrastructure. Routing schemes such as Dynamic Source Routing [2] and Ad Hoc On-Demand Distance Vector [3] were implemented to perform basic routing operation like forward data packets from a source to a destination. Routing schemes should consider the characteristics of the MANET. The prime needs of a MANET are reliability of data transmission, multipath selection [4], and providing security [5] which increases the network performance. Many of the researches have been developed to achieve this goal. On-demand routing [6] is one of the essential functions in MANET. The routing schemes [7] should be reliable, robust, and flexible in an ad hoc environment. Routing function is restricted by dynamic topology and link failure of the nodes. The mobility of the nodes increases the complexity of routing function because it causes of link failure between nodes. This frequent link failure leads to routing overhead and topology management, reduces the reliability of data transmission, and reduces the efficiency of the network. Hence, the link failure in MANET becomes a vital issue. Further, this sort of link failure also leads to frequent path failures. As a result, the reliability of data transmission gets reduced; the lesser the packet delivery ratio, the longer the end-to-end delay. Retransmission of data packets in a MANET is costly increasing the control message overhead and reducing the efficiency of the routing function. Hence, it is very much essential to select an alternate path or to form multiple paths for ensuring reliability of data transfer when a link failure occurs. And also loop-free path is very much interested in finding an optimal path among multiple paths between source and destination in a network.

Multipath routing in a MANET [8] is established in order to increase the reliability of data transmission that provides load balancing among the nodes. The use of multiple disjoint paths transferred the data in parallel that significantly increases the packet delivery ratio. Multipath routing schemes [9] deal with the problem of scalability, confidentiality, integrity, and network lifetime. Multiple-path routing [10] between source and destination ensures reliability of the data transmission in a MANET. Existing multipath routing schemes in a MANET lead to problems such as flooding, empty set of neighbors, flat addressing, widely distributed information, large energy consumption, interference, and load balancing issues. Therefore, the efficient multipath routing scheme is proposed to solve one or more of these issues. And also the existing multipath routing schemes do not perform well in dynamic environment change and frequent path failure. They also generate a routing overhead in the network. The routing overhead occupies a considerable portion of network bandwidth and the energy of the mobile node exhausts rapidly. Hence, with minimum overhead, the reliable multipath routing protocol is essential for designing to restrict the participation of mobile nodes in a route discovery phase that ensure reliability of data transmission. Evolutionary mechanism paradigm [11, 12] is most suitable to resolve multiobject problems because they are based on population. It generates a set of solutions in one run. There is no single solution that can be termed as the optimal solution [13] in multiobjective problems. In particle swarm optimization (PSO) [14, 15], the potential solutions of the problem are called particles. The group of particles becomes a swarm that searches for an optimum solution. Particle swarm optimization [16, 17] is a stochastic optimization technique in which the particles fly in the search space and adjust to the velocities dynamically according to their historical behaviors. This process guides the particles to soar toward the better search area in the search space [18]. It is difficult to ensure the reliability of the data transmission from source to destination because of its dynamism property and frequent link failure in MANET. The PSO can be applied to solve this kind of issue.

We design an energy-aware multipath routing scheme based on particle swarm optimization that uses continuous time recurrent neural network to solve optimization problems. Its goal is to discover multiple loop-free paths by using PSO technique. For gaps in the above literatures review, this work gives an optimum path selection technique that produces a set of optimal paths between source and destination to enhance the reliability of the data transmission.

The remainder of the paper contains five sections. Section 2 summarizes previous research works. Section 3 describes the proposed energy-aware multipath routing scheme based on particle swarm optimization technique in a MANET. Section 4 presents our simulation results and a relevant performance analysis. Finally, Section 5 presents our conclusions and future direction.

#### 2. Related Works

Evolutionary mechanisms [19] and clustering scheme [20] play a vital role to find optimal solutions for network partition. A number of evolutionary mechanisms have been proposed such as genetic algorithm [21, 22], artificial neural network system [23], particle swarm intelligence [24], and particle swarm optimization based clustering [25–27] that reveals the best global solutions. Dengiz et al. [28] proposed a particle swarm optimization algorithm that conceptualizes an autonomous topology optimization for mobile ad hoc networks using multiple mobile agents. The MANET communication is represented as network flows and optimization using a maximum flow model. This representation is very responsive to small changes in topology when evaluating network connectivity. Goswami et al. [29] proposed a fuzzy ant colony based routing protocol using fuzzy logic and swarm intelligence to select the optimal path by considering optimization of multiple objectives while retaining the advantages of swarm based intelligence algorithm. Ali et al. [30] proposed a multiobjective solution by using multiobjective particle swarm optimization algorithm to optimize the number of clusters in an ad hoc network and reduce the network traffic. This scheme performs intercluster and intracluster traffic that is managed by the cluster-heads. The authors considered the performance measures such as degree of nodes and power consumption of the mobile nodes.

Nasab et al. [31] proposed a multicast routing based on the particle swarm optimization (MPSO) that focused on efficient energy consumption and delay in multicast routing in MANET. It selects the node with the minimum energy consumption in the route selection and builds a multicast tree with minimum delay. There exists route failure in all route discovery methods resulting in data loss and routing overheads. Manickavelu and Vaidyanathan [32] proposed a Particle Swarm Optimization based Lifetime Prediction (PSOLP) algorithm for route recovery in a MANET. This scheme predicts the lifetime of link and node in the available bandwidth based on the parameters such as relative mobility of nodes and energy drain rate. The parameters are fuzzified and fuzzy rules have been formed to decide on the node status for prediction. Then, this piece of information is exchanged among all the nodes for verifying status of a node before data transmission.

Yun-Sheng et al. [33] proposed a multicast QoS based routing approach using genetic algorithm. It uses the available resources and minimum computation time in a dynamic environment. This scheme optimizes the routes by selecting the appropriate values for genetic operations like crossover, mutation, and population size. Radi et al. [34] proposed Low-Interference Energy-efficient Multipath Routing protocol (LIEMRO) to distribute source node traffic over the established paths. It also provides load balancing that estimates the optimal traffic rate of the paths. LIEMRO starts packet transmission immediately after the first path is established. Whenever a new path is created, the load balancing algorithm redistributes the source node traffic, according to the relative quality of the paths. Hurni and Braun [35] proposed a multipath routing protocol based on AOMDV. The objectives of these multipath schemes are to improve energy efficiency and reduce latency through load balancing and using cross-layer information. In order to reduce the end-to-end delay of data forwarding, each node utilizes the information provided by the MAC layer to transmit its packets to the neighboring node that wakes up earlier. Ghiasi and Karimi [36] proposed an algorithm of learning automata adjusting learning rate on neural network. It is a combination of the back-propagation algorithm, a local search algorithm, and learning automata to provide efficient global search. Mobile network parameters were measured for training and testing the neural network. The learning automata approach does not find optimal solution when the number of nodes increased in the network.

#### 3. Energy-Aware Multipath Routing Scheme Based on Particle Swarm Optimization

MANET consists of mobile nodes with limited energy and wireless link. Each mobile node forwards the packets from source to destination. MANET is represented as directed graph . The vertices are a symbol of the mobile nodes and the neighbor node. An edge is a symbol of a wireless link between nodes , which forward packets to others. The energy consumption for forwarding packets from a node to node is given by where is number bits and is the distance between nodes. , are energy dissipated per bit to forward and receive packets, respectively. Energy consumption for receiver is calculated by The proposed EMPSO routing scheme is composed of three phases: (i) route setup phase, (ii) route discovery phase, and (iii) route maintenance phase. In the route setup phase, each node acquires its metadata of the neighborhood. This metadata is used in the route discovery to find the best next-hop node towards the destination node. The route discovery is activated whenever a source wants to transmit data to destination in an on-demand fashion that prevents multiple interference between source and destination. The route maintenance phase handles path failures during data transmission.

##### 3.1. Route Setup Phase

In the route setup phase, source node initiates a data transmission for forwarding packets to the destination. Each node in a MANET obtains its metadata of the neighborhood, which also includes the transmission cost () of its neighbors towards the destination node. The value of a link indicates the required number of transmissions for a successful packet reception at the receiver. The transmission cost of a link is given as follows:where and are the probabilities of forward and backward packet reception over a link, respectively. In the initialization phase, each node broadcasts the control packets and stores the number of successfully received packets from its neighbors in the routing neighborhood table. Then, the destination node sets its transmission cost to zero and broadcasts this value to its neighbors, when a node receives a transmission cost included in a packet.

##### 3.2. Route Discovery Phase

Whenever a source node wants to transmit data to destination, the route discovery phase is initiated to find multiple paths from the source to destination. The proposed multipath routing protocol uses reliability measures such as transmission cost, optimal traffic ratio, and remaining energy. The source node starts the route discovery by transmitting a route request packet (RR) towards the destination node. Whenever an intermediate node receives a RR packet, it computes the transmission cost, optimal traffic ratio, and remaining energy for a path that is established between the source and the destination. Then, it also used a found path to forward the RR packet to the neighboring node with minimum cost. The reliability measures are stored in the routing table of a node in MANET. The proposed EMPSO scheme uses a continuous time recurrent neural network to find an optimal path among multiple paths. CTRNNs are more computationally efficient in order to use a system of ordinary differential equations to model. The three weight factors such as transmission cost, energy factor, and optimal traffic ratio are taken into the account in CTRNN to find an optimal path. The weight factor of transmission cost is given inwhere is the weight factor for transmission cost, is the control packet transmission ratio and is the data packet transmission ratio from node to node , respectively, and is the time taken for and . The weight factor for optimal path ratio and remaining energy of a node are calculated as inFor a neuron in the network with action potential the rate of change of activation is given in Notations used in (5) and (6) are as follows: : th path; : total energy required for packet forwarding from node to node ; : transmitted energy, received energy, and ideal energy of a node, respectively; : time constant of postsynaptic node; : rate of change of activation of postsynaptic node; : weight vector of transmission cost from presynaptic to postsynaptic node; : weight vector of optimal path ratio from presynaptic to postsynaptic node; : weight vector of remaining energy of node; : sigmoid of ; : bias of presynaptic node; : input to node.

Figure 1 shows the flow diagram of the proposed scheme. Initially, multiple paths are found by route discovery phase. Then, reliability measures for a path are calculated with help of CRRNN. An optimization technique called PSO is used to find better link quality in a path. PSO can be applied to optimization problems that are partially in dynamic topology changing environment. PSO is an evolutionary optimization technique that may be used to seek a good set of weights in CTRNN. PSO is applied to find the best nodes (particles) involved in a path. PSO is metaheuristic that searches large spaces of candidate solutions. A route with a better link quality is selected for forwarding data from source to destination. If a better link quality is not found, PSO function is performed again until global best solution has been found. PSO reduces the traffic and routing overhead of the optimization process and finds the node with best link quality in an ad hoc network.