Article Preview
TopIntroduction
IoT is a set of networks consisting of a number of sensors as well as actuator devices (Atzori et al., 2010). Sensors (devices) are deployed to sense the data from the environment, and actuator devices act as agents that initiate the actions. These two together made smart applications viz. smart wearable, smart grids, security surveillance, smart business, health care, smart home, mobile ticketing, smart museum, robot taxi, smart tracking (Al-Karaki & Kamal, 2004), landslide detection, fire detection, smart agriculture (Poluru & Kumar, 2019), etc. All the IoT devices have limitations such as energy, memory, and computation. In general, IoT devices are capable of configuring an ad-hoc network autonomously and transfer the collected data among them. Also, a few IoT devices act as gateways that are connected to the internet to communicate the data to remote locations (Xu et al., 2017). Thus, energy-aware protocols should be lightweight and simple, which consumes less energy. The main factor that decreases energy in IoT networks markedly is wireless communications. The restricted energy of IoT devices raises the objective of preserving energy levels for an extended lifetime of the IoT network (Anastasi et al., 2009; Li et al., 2017).
Energy efficiency plays an anchor role (Azharuddin & Jana, 2017) in large-scale IoT due to extended geographical distances for data transfer, an ample number of sensors and huge quantity of data to be sensed. Node clustering-based hierarchical routing is a well-known class of data sensing techniques. Here, each unit of the autonomous network is called as clusters (Kumar et al., 2018; Mostafaei et al., 2017). Each cluster consists of a group of sensor nodes that communicate data among them. For consistency, a CH (Chatei et al., 2017; Sirdeshpande & Udupi, 2017) is being selected, which is responsible for transmitting the data down to all the nodes in the cluster to establish a reliable network. CH can be selected either dynamically based on residual energy, distance etc., or pre-assigned CH statically by the network designers. In the literature, researchers studied extensively and developed various algorithms on this issue in the IoT networks. Clustering structures are employed to alleviate the number of energy holes being created because of a non-reliable network structure. Depending on the various tasks nodes perform, they are categorized into CH as well as CMs (Kannan & Raja, 2015). CH takes the responsibility of collecting the information from the surrounding environment as well as forwards it to the BS. Thus the information is transferred in the cluster is diminished as well as energy consumption. The way CH is being selected (Ansari & Cho, 2018; Priyadarshini & Sivakumar, 2018) directly impacts the network lifetime and also decreases the energy consumption (Selbach, 2013) during cluster formation as well as reselection. Though many non-probabilistic approaches were employed in the literature for improving energy-efficient routing algorithms, it is continued to exist as an open issue in the IoT environment (Mohanty & Kabat, 2016). The cluster-based IoT architecture (Poluru & Kumar, 2019) as shown in Figure 1.
The main intention of this paper is to design a multi-objective optimization (Mirjalili et al., 2017) approach to handle the above-said clustering and routing problems in IoT. Thus, a Multi-Objective Ant Lion Optimizer meta-heuristic Algorithm (MOALOA) is developed to achieve efficient clustering and fast routing functionalities such as energy efficiency, throughput, extended network lifetime and hign convergence compared with ELEACH, mACO, MFO-ALO, and ALOC. The other multi-objective meta-heuristic algorithms developed in the literature are MOPSO (Xue et al., 2012), MOGWO (Mirjalili et al., 2016), Hybrid algorithms (Gadekallu & Khare, 2017; Reddy & Khare, 2017; Reddy & Khare, 2018; Reddy et al., 2019), multi-objective Bee Algorithm (Akbari et al., 2012), and multi-objective Bat Algorithm, etc. However, the MOALOA has gained advantages over the above-said algorithms in the following ways: (1) It fits into large-scale IoT networks easily. (2) It generates quick optimal solutions to avoid local optima. (3) It converges very fastly.
The above said challenges in designing efficient routing protocols motivated the work contributed in this paper. This paper proposed a meta-heuristic based MOALO algorithm that achieves the following objectives: