Time Window-Based On-Demand Scheduling For Multiple Mobile Chargers

With recent advancement in techniques such as wireless communication,sensor node,microelectronics,Wireless Sensor Networks(WSNs)are used in many fields.However,since sensors are powered by energy-constrained batteries in wireless sensor networks,the operational time of networks usually is limited,which seriously hinders the development of sensor networks.With the help of recent breakthrough in wireless energy transfer,the deployment of mobile charger with wireless transmission devices in the networks to construct Wireless Rechargeable Sensor Networks(WRSNs),it has become a promising solution for prolong network lifetimes.The traditional research on mobile charger scheduling is usually aimed at small-scale networks scenarios,i.e.,single mobile charger is used to wireless supplement energy for sensor nodes in the network.Nevertheless,with the gradual expansion of the application range of WRSNs,a single mobile charger is difficult to meet the charging demands of sensors in large-scale network scenarios.Thus,this paper studies that the scheduling problem of multiple mobile chargers for large-scale network scenarios.In large-scale WRSNs,it is not only necessary to dispatch multiple mobile chargers to replenish energy for a large number of sensor nodes with charging demands,but also the charging demands of these nodes vary greatly and dynamically,which brings many challenges to design an efficient multi-mobile charger scheduling scheme.Thence,this paper conducts research on the on-demand charging scheduling problem of multiple chargers in large-scale network scenarios.The main contributions include the following three points:Firstly,the time window is introduced to model the actual charging demands of all sensors.When the remaining lifetime of the sensor is lower than a certain threshold,it’s time window will be opened,and when the energy of the sensor is exhausted,the time window will be closed.In addition,in order to ensure the survival of the all sensors,the mobile chargers needs to replenish energy for the sensor while the time window is open.Thus,the time window can be dynamically changed according to the actual energy consumption rate of the sensors,ensuring that the designed charging scheduling scheme can reasonably arrange the charging sequence and avoid the death of sensor nodes.Secondly,this paper proposes a multi-charger on-demand charging scheduling based on hybrid population incremental learning.First of all,the scheme establishes a three-dimensional probability matrix describing the path distribution according to the number of mobile chargers.Then,this paper designs a probabilistic space-time selection function,which is used as the judgment basis for the mobile charging vehicle to select sensors in the network to join the charging path each time.Furthermore,considering that the probabilistic space-time selection function is essentially a greedy strategy,each time the sensor node with the largest value is selected to be added to the path,this paper adds a random operator to the probabilistic spatiotemporal selection function and defines it as a hybrid probabilistic spatiotemporal selection function.Finally,the hybrid probabilistic spatiotemporal selection function is used to generate the optimal charging paths for multiple mobile chargers.Finally,this paper compares the proposed scheme with the existing charging scheme through experiments,and the proposed scheme has better results.On average,the multi-charger on-demand charging scheduling based on hybrid population incremental learning in this paper has 17.8%,48.7% and 56% performance advantages over the three charging scheduling schemes of m TS,NJNP,and Random,which means that the proposed scheme can better utilize the energy resources of the mobile charger to charge more sensors in the network.