Multi-objective particle swarm optimization based rendezvous point selection for the energy and delay efficient networked wireless sensor data acquisition

Traditionally, the non-rendezvous points transmit data towards Rendezvous Points (RPs), and Mobile Sink (MS) visits RPs to collect data. The existing body of research mitigates the problem of data acquisition latency, load on RPs, and energy consumption by regulating the number of RPs. Fewer RPs benefit the data acquisition latency, whereas increased RPs benefit the multi-hop forwarding and load on RPs. This paper takes up these issues and models as multi-objective optimization problem attempting to minimize data collection latency, data load among RPs, and the number of RPs. Particle Swarm Optimization (PSO) is the widely used meta-heuristic method to solve a multi-objective optimization problem with better convergence and minimum overhead. This paper introduces a Multi-Objective Particle Swarm Optimization based RPs Selection (MOPSO-RPS) method for energy and delay efficient data collection. MOPSO-RPS applies a new encoding scheme to generate variable dimension particles that represent each possible set of RPs. Additionally, a new inertia weight tuner is also referred to enhance the convergence speed of multi-objective PSO towards the optimal solution. However, it might happen that after updating the location and velocity in each iteration, the particles become invalid due to the violation of the search space boundary. Thus it adopts a valid particle generator to create valid particles. Moreover, an improved ant colony optimization is also applied to construct the trajectory of MS with fast convergence speed towards the optimal solution. Simulation results depict that the proposed MOPSO-RPS result in 18.61%, 21.11%, and 10.71% average improvement in energy consumption, data load among RPs, and data acquisition latency, respectively, for different number of sensor nodes when compared with the state-of-the-art methods. [Display omitted] •MOPSO-based RPs selection to minimize data acquisition latency, data load of RPs and number of RPs.•An efficient encoding scheme to generate variable dimension particles that represent a set of RPs.•An adaptive tuning of inertia weight to enhance the convergence speed towards the optimal solution.•Improved ACO-based delay efficient trajectory formation for data acquisition.