无线供能移动边缘计算（wireless powered-mobile edge computing, WP-MEC）集成了移动边缘计算和无线功率传输技术，旨在解决移动设备计算能力不足和持续能源供应问题。然而由于WP-MEC中不同边缘服务器供电能力和计算能力不同、移动设备需执行任务的延迟忍耐时间异构，以及移动设备与服务器之间时变的无线信道给系统时间资源分配和任务处理带来了巨大挑战。基于此，从WP-MEC网络的异构服务器选择、计算卸载和资源分配联合优化的角度开展研究，为提高系统有效计算率，提出了基于延迟敏感性任务加权平均的坐标下降（joint optimization scheduling algorithm with weighted average of delay-sensitive tasks and coordinate descent, WADT_CD）联合调度算法。首先，综合考虑时变无线信道增益、异构任务延迟、异构边缘服务器的发射功率和计算能力，设计基于延迟敏感性任务加权平均（scheme of weighted average of delay-sensitive tasks, WADT）的异构服务器选择策略。其次，考虑WP-MEC网络模型特性，设计基于一维时间变量二分搜索的坐标下降（method of coordinate descent, CD）算法解决移动设备卸载决策和时间资源分配问题。最后，通过仿真实验与多种算法进行对比，验证了所提方法的优越性，并且分析了在不同规模边缘设备、异构任务比例时所提算法的有效性。

WP-MEC (wireless powered-mobile edge computing) combines mobile edge computing and wireless power transfer technologies. It aims to address the challenges of insufficient computing power and continuous energy supply for mobile devices. However, the heterogeneous power supply and computational capabilities of different edge servers in WP-MEC, as well as the varying delay tolerance times of mobile devices and the time-varying wireless channels between devices and servers, create significant challenges in system resource allocation and task processing. To address these challenges, research has been conducted from the perspective of joint optimization of heterogeneous server selection, computation offloading, and resource allocation in WP-MEC networks. In order to improve the effective computation rate of the system, WADT_CD(weighted average of delay-sensitive tasks-based coordinate descent) joint scheduling algorithm was proposed. First, a heterogeneous server selection strategy was designed based on WADT(weighted average of delay-sensitive tasks), taking into account the time-varying wireless channel gain, heterogeneous task latency, transmit power, and computational capability of edge servers. Second, CD (coordinate descent) algorithm was developed, based on one-dimensional time-variable binary search, to solve the mobile device offloading decision and time resource allocation problems. Finally, the proposed method was validated through simulation experiments, which compare its performance with various algorithms, and the effectiveness of the proposed algorithm was also analyzed at different scales of edge devices and heterogeneous task ratios.