To make a good balance between performance, cost, and power consumption, a hybrid intelligent reflecting surface (IRS)-aided directional modulation (DM) network is investigated in this paper, where the hybrid IRS consists of passive and active reflecting elements. To maximize the achievable rate, two optimization algorithms, called maximum signal-to-noise ratio (SNR)-fractional programming (FP) (Max-SNR-FP) and maximum SNR-equal amplitude reflecting (EAR) (Max-SNR-EAR), are proposed to jointly design the beamforming vector and phase shift matrix (PSM) of hybrid IRS by alternately optimizing one and giving another. The former employs the successive convex approximation and FP methods to derive the beamforming vector and hybrid IRS PSM, while the latter adopts the maximum signal-to-leakage-noise ratio method and the criteria of phase alignment and EAR to design them. Simulation results show that the rates harvested by the proposed two methods are slightly lower than those of active IRS with higher power consumption, which are 35 percent higher than those of no IRS and random phase IRS, while passive IRS achieves only about 17 percent rate gain over the latter. Moreover, compared to Max-SNR-FP, the proposed Max-SNR-EAR method makes an obvious complexity degradation at the price of a slight performance loss.

翻译：为在性能、成本和功耗之间取得良好平衡，本文研究了一种混合智能反射面（IRS）辅助的定向调制（DM）网络，其中混合IRS由无源和有源反射单元组成。为最大化可达速率，提出了两种优化算法：最大信噪比-分数规划（Max-SNR-FP）和最大信噪比-等幅反射（Max-SNR-EAR），通过交替优化波束成形向量和混合IRS的相位偏移矩阵（PSM）实现联合设计。前者采用连续凸近似和分数规划方法推导波束成形向量及混合IRS相位偏移矩阵，而后者则采用最大信漏噪比方法结合相位对齐与等幅反射准则进行设计。仿真结果表明：所提两种方法获得的速率略低于功耗更高的有源IRS，但比无IRS和随机相位IRS高出35%，而无源IRS仅比后者获得约17%的速率增益。此外，与Max-SNR-FP相比，所提Max-SNR-EAR方法在轻微性能损失代价下显著降低了算法复杂度。