Determining the trade-offs between data delivery and energy consumption in large-scale WSNs by multi-objective evolutionary optimization

Abstract

Autonomous Wireless Sensor Networks (WSNs) form the basis of several Internet of Things (IoT) applications. Efficiency in routing data is a major concern in these networks since sensor nodes are energy-constrained and mostly unaware of the optimal routes regarding a set of QoS metrics. In this paper, the energy consumption and data delivery reliability are deemed the most important factors to the adequate overall network and application operation. Nevertheless, to enhance the chances of data packets to arrive at destinations, in general, energy consumption and reliability have to compromise. Although many routing techniques exist for WSNs, it is hard to determine, in general, if they really can find good solutions considering such trade-off. In light of this, we present a multi-objective integer problem (MOIP) to the routing problem in WSNs to optimize both energy consumption and delivery reliability. Moreover, to solve the proposed MOIP, we introduce an efficient evolutionary algorithm based on the Non-dominated Sorting Genetic Algorithm II (NSGA-II). The solution method is capable of constructing Pareto set approximations in a low computational time that can be used to better evaluate the solutions provided by routing protocols prior to deployment. The results of experiments with medium-sized and large-scale scenarios based on real case studies indicated the clear compromise between energy efficiency and delivery ratio. Moreover, by the results achieved by NSGA-II, the scalability does not affect the performance of the routing in optimizing the two objectives.