Beamforming Design with Bilevel Optimization for RIS-Assisted SWIPT Systems

Simultaneous wireless information and power transfer (SWIPT) has been regarded as one of the most promising technologies for the Internet of Things (IoT). However, in some scenarios where one function takes priority over the other one, existing single-level optimization models do not apply to formulate this class of problems. To address it, we employ bilevel optimization in this paper and consider a reconfigurable intelligent surface (RIS)-assisted SWIPT system. Specifically, the harvested power is maximized in the lower-level problem and the sum rate is maximized in the upper-level problem. We first derive the optimal set of the lower-level problem and then transform the bilevel problem into an equivalent but more tractable form. Subsequently, a block-wise bilevel-based beamforming design (B
³
D) algorithm is developed, which converges to the stationary point of the bilevel problem. Finally, simulation results are presented to verify the performance of the proposed algorithm. Compared to the existing optimization schemes, the proposed algorithm achieves a higher sum rate with the same harvested power.