基于改进TD3的MEC多任务计算卸载

在多用户多任务场景下, 使用传统的决策算法去对短时间内接踵而来的任务进行计算卸载决策, 已经不能满足用户对决策效率和资源利用率的要求. 因此有研究提出使用深度强化学习算法来进行卸载决策以满足各种场景下的需求, 但是这些算法大多只考虑卸载优先的策略, 这种策略使用户设备(UE)被大量闲置. 我们提高了移动边缘计算(MEC)服务器和用户设备(UE)的资源利用率, 降低计算卸载的错误率, 提出了一种本地优先和改进TD3(twin delayed deep deterministic policy gradient)算法相结合的决策卸载模型, 并设计了仿真实验, 通过实验证明该模型确实可以提高MEC服务器和UE的资源利用率并降低错误率.     

PVP修饰石墨烯对SBS改性沥青流变性能的影响

为研究经聚乙烯吡咯烷酮(PVP)处理后的石墨烯对SBS改性沥青流变性能的影响,通过高速剪切机在SBS改性沥青中掺入PVP修饰石墨烯,制备了石墨烯-SBS复合改性沥青,并采用动态剪切流变仪、温度扫描和线性振幅扫描试验,对复合改性沥青流变性能及抗疲劳性能变化规律进行研究,同时结合弯曲梁流变仪试验探索了复合改性沥青低温抗裂性...     

Structure and microwave dielectric properties of Li(Al1-xLix)SiO4-x ceramics

Li(Al_1-xLi_x)SiO_4-x (x = 0.005, 0.01, 0.015, and 0.02) ceramics were synthesized via a traditional solid phase reaction method with different sintering temperatures. To determine the positions occupied by Li⁺ in the lattice, the defect formation energies and total energies of various sites of LiAlSiO₄ (LAS) occupied by Li⁺ were examined, and the energy of LAS systems were calculated using density functional theory of first-principle with the CASTEP module. The results demonstrated that the Al-sites occupied by Li⁺ had the lowest formation energies and total energy, so Li ⁺ should substitute Al³⁺. The impacts of replacing Al³⁺ with Li⁺ on the bulk density, sintering properties, phase composition, microstructure, and microwave dielectric properties of Li(Al_1-xLi_x)SiO_4-x (0 = x ≤ 0.02) ceramics were thoroughly studied. With Li⁺-doping, the sintering temperature decreased from 1300 °C (x = 0) to 1175 °C (x = 0.02), while the Q × f and τ_f values of LAS ceramics significantly increased. The Li(Al_0.99Li_0.01)SiO_3.99 ceramic was fully sintered at 1250 °C for 10 h to obtain excellent microwave dielectric properties: ε_r = 3.49, Q × f = 51,358 GHz, and τ_f = ?51.48 × 10^?6 °C^?1.