Differential Evolution-Based Trajectory Planning (DEBTP) in Intermittently Connected Delay Tolerant Wireless Sensor Networks

The Internet of Things-based Intermittently Connected Delay-Tolerant Wireless Sensor Network (IoT-based ICDT-WSN) comprises a set of WSNs wherein the nodes belonging to two different WSNs cannot communicate with one another. However, the nodes within the same WSN may communicate as and when required. Thereby, the data collection in an IoT-based ICDT-WSN is a challenging task due to the intrinsic characteristics of the network in addition to those of the participating nodes. Like, the nodes are resource-constrained in nature and the network suffers from the absence of always-on connectivity as compared to its traditional counterpart. To facilitate the data collection in such a network, a mobile agent is deployed which visits each of the constituting WSNs to collect their respective data and the collected data is then reported to the base station by the mobile agent. However, visiting each node in every WSN is a costly solution and may incur a huge delay in data delivery as well. Therefore, the network designates a representative (rendezvous) node for each of the participating WSNs and the mobile agent visits only these representative nodes. The proposed work selects appropriate rendezvous nodes for each participating subnetwork and defines a movement trajectory for the mobile agent too in a single phase. The proposed scheme targets to deliver the data packets collected from the constituting subnetworks to the base station in the least possible time. It employs the differential evolution (DE) metaheuristic scheme for the selection of rendezvous nodes and trajectory definition of the mobile agent. The scheme utilizes a convex-hull algorithm while considering the spatial distribution of the sensor nodes in the network to decide upon the most feasible path for the mobile agent. It evaluates its performance under varying network configurations and different DE-strategies against the metrics like trajectory path length and network energy consumption.