Wireless Rechargeable Sensor Networks based on wireless energy transmission successfully solve the problem of node death caused by energy shortage in traditional wireless sensor networks. Existing studies on joint energy replenishment and data collection have planned charging path first and then perform data collection. This approach does not effectively extend network life and improve network performance in data collection. In this paper, we jointly consider the impact of energy replenishment and data collection on network performance and use the Least Squares Support Vector Machine based regression prediction with dynamically changing energy consumption rate and data generation rate of sensor nodes. To solve the above problem, we propose a multi-objective path planning model for joint energy replenishment and data collection with the optimization objectives of maximizing the remaining lifetime of sensor nodes, maximizing the data collection of MC, and minimizing the amount of data loss. The conventional exact algorithm is difficult to solve the path planning problem, so heuristic algorithms have gradually become the main algorithm to solve the problem. Based on discrete fireworks algorithm and co-evolutionary algorithm, we present a grid-based multi-objective cooperation fireworks algorithm. Simulation results show that the proposed algorithm performs better in convergence and diversity. In terms of the average remaining lifetime of sensor nodes, the proposed algorithm is 0.43%, 0.52%, 1.97%, and 0.81% higher than MPACO, MODFA, NSGA-II, and SPEA-II, respectively. Similarly, the target value of solutions in data collection increased by 1.87%, 1.22%, 4.49%, and 2.10%, respectively.

基于无线能量传输的无线可充电传感器网络成功地解决了传统无线传感器网络因能量不足导致的节点死亡问题。现有的联合能量补充和数据采集研究都是先规划充电路径，再进行数据采集。这种方式并不能有效地延长网络寿命和提高数据采集中的网络性能。在本文中，我们共同考虑能量补充和数据收集对网络性能的影响，并使用基于最小二乘支持向量机的回归预测与动态变化的能量消耗率和传感器节点的数据生成率。为解决上述问题，我们提出了一种用于联合能量补充和数据收集的多目标路径规划模型，其优化目标是最大化传感器节点的剩余寿命，最大化MC的数据收集，以及最小化数据丢失量。传统的精确算法难以解决路径规划问题，启发式算法逐渐成为解决该问题的主要算法。基于离散烟花算法和协同进化算法，我们提出了一种基于网格的多目标协作烟花算法。仿真结果表明，该算法在收敛性和分集性方面表现较好。在传感器节点的平均剩余寿命方面，该算法分别比 MPACO、MODFA、NSGA-II 和 SPEA-II 高 0.43%、0.52%、1.97% 和 0.81%。