Proximal Policy Optimization for Energy-Efficient MEC Systems with STAR-RIS Assistance

The growing popularity of Internet of Things (IoT) devices has led to an escalating demand for efficient data processing and transmission solutions. The concept of Mobile Edge Computing (MEC) has emerged as a potential solution to tackle these challenges by bringing computation closer to IoT devices. Nevertheless, the establishment of reliable communication connections between IoT devices and MEC servers continues to be a significant issue, especially in situations where achieving line-of-sight (LOS) conditions is troublesome. This paper studies simultaneously transmitting and reflecting reconfigurable intelligent surfaces" (STAR-RIS) to enhance communication links in MEC environments. STAR-RIS leverages the capabilities of conventional RIS to simultaneously transmit and reflect signals, thereby providing 360
^°
coverage. We formulate the energy minimization for all IoT devices in the STAR-RIS-assisted MEC system by jointly optimizing the energy-efficient offloading, amplitude, and phase shift coefficients of reflection and transmission of STAR-RIS elements and power control. Due to the non-convexity and coupling variables, the proximal policy optimization (PPO) technique has been adopted as a viable solution. The experimental findings presented in this study provide evidence of the efficacy of our suggested algorithm in comparison to the benchmark schemes.