Article Preview
TopIntroduction
According to Zakrzewski et al, (2009), and Rani and Gulati (2012) The efficient energy utilization in WSN is a bottleneck problem that affects the network's performance and lifetime. Researchers recently addressed energy consumption attentiveness and power management approaches to tackle this problem. Optimal routing techniques and energy optimization usage significantly affect the WSN performance and guarantee the extension of the network lifetime. Due to WSN constraints and especially the sensors` energy scarcity, a smart routing should be done to balance the energy consumption among nodes. Therefore, prolong the network lifetime and ensuring network coverage. Deploying intelligent and intellectual techniques enhances the effectiveness of wireless sensor networks. Researchers have studied and examined different soft computing paradigms to optimize WSN routing with the consideration of the power consumption, network challenges, and design and deployment aspects. The soft computing paradigms such as RL, SI, EA, FL, NN, and AIS have been applied to different WSN applications and deployment based on their other characteristics (Swami et al, 2013). The section below in this survey work summarizes the recent implementations of soft computing paradigms in routing in WSN with its dynamical and heterogeneous characteristics.
Background
The two efficient algorithms in a wireless sensor network are AODV and DSR. On-demand routing protocol AODV establishes the route only when there is a requirement. This protocol uses two different broadcast messages such as RREQ and RREP. The advantages of AODV over DSR are that they require less storage. The hello messages will be sent to the network to see whether the nodes are active. If there is no response to the message, then it is considered that the links are broken. Due to which communication overhead is formed in the network (Tuteja et al, 2010). Dynamic source routing will have all the details included in the packet header. It is also a demand-driven protocol. These protocols use route detection and conservation. The network is formed via sending the route request and when the packets remain not sent, a fault message is sent to entire nodes in the network. The main variance between these two algorithms is a direction-finding feature. The novel approach will consider high storage space with less communication overhead. These improve the time of the network activated (Johnson et al, 2001; Mohapatra and Kanungo, 2012).
Table 1.
Comparison of AODV and DSR
| Scenarios | AODV | DSR |
| Demand-driven | Yes | Yes |
| Source routing | Computes and Updates the route | Route detection and maintenance adds a sequence number Multipath. |
| Routing load | Comparatively high | Less |
| Route overhead | Find routes in Caches hence reduces | More compared to DSR because there is no caching used |
| Standardized MAC load | Less when Compared to Lower pause time | Relatively High |
| Performance | Higher when it is in complex mobility scenarios | Comparatively less. |
Table 1 compares AODV and DSR. In AODV the hello message will be sent when there is a demand of creating the network. RREQ will be sent to all the nodes in the network. This request will contain all the details in the network. The sequence number will help us know the destination address of the packet to be sent. RREP is the route reply from the node if the packet reaches the destination node (Lee et al, 2000; Chakeres et al, 2004; Sethuraman and Kannan, 2017). DSR protocol needs an extra storage capacity since it uses caching technology. Here the route detection is initialized when a node needs a route to be established in the network. Flooding technology is used in this protocol. All the packets sent in the network will have all details regarding the routing of data. In case of a link failure, RERR is sent among the source. Here energy is not considered as a factor (Hu and Johnson, 2000).