An energy aware grouping memetic algorithm to schedule the sensing activity in WSNs-based IoT for smart cities

Abstract

Wireless Sensor Networks (WSNs) are the main component in the Internet of Things (IoT) and smart cities to sense our environment, gather essential and meaningful data, and forward it to the base station (BS). Nowadays, IoT utilizes WSN as a necessary platform for sensing and communication of the data. One of the main strategies to minimize consumption of energy in a WSN with highly dense sensors is to maximize sleeping sensors at each time. This way implies to schedule the activity of sensors, i.e. determining when a sensor node is kept idle (sleep mode) and when a sensor node is activated to sense environment (active mode). Due to heterogeneity of the sensor nodes in WSNs-based IoT for smart cities, one approach for scheduling the sensing activity is clustering the sensors into K mutually different subsets, so that every subset of sensors alone can cover all targets of the network. In this case, we can solve finding the maximum number of sensor subsets, or equivalently sensor covers problem, by conversion it to SET K-COVER problem. In this paper, an energy aware Grouping Memetic Algorithm (GMA) is proposed for solving the SET K-COVER problem. The proposed GMA varies in four general ways from any of the other evolutionary algorithms for solving the SET K-COVER problem. First, to transform solution structures linked to the SET K-COVER problem into chromosome genes, a new encoding scheme was being used. Second, specific genetic operators appropriate for the chromosomes are used, based on the encoding used. Third, to direct the search process in the solution space of the SET K-COVER problem, a novel proper fitness function is proposed. Fourth, a local improvement algorithm which uses a sensor dominance rule is proposed. This study, carried out detailed experiments of different numbers of targets and different numbers of sensors in WSNs to assess the proposed algorithm. In several instances of the problem, the experiments demonstrate that the algorithm works considerably better in terms of solution quality than many other heuristics and evolutionary algorithms and significantly outperformed the evolutionary algorithms in terms of runtime, suggesting the usefulness of the proposed algorithm to increase the lifetime of WSN in IoT and smart cities.