Computation Efficiency Optimization for Active-RIS-Assisted Mobile Edge Computing Systems

This paper investigates an active reconfigurable intelligent surface (ARIS) assisted uplink mobile edge computing (MEC) systems. We propose an ARIS-aided user tasks computing scheme called ARIS-UTC, where a user simultaneously performs local computing and offloading computing to a multi-antenna access point (AP) assisted by an ARIS. To strike a balance between the data volume of computing tasks and energy consumption, we formulate a computation efficiency (CE) maximization problem for the proposed ARIS-UTC by jointly optimizing the user's transmit power, local computing frequency, ARIS reflection coefficients, and the receive beamforming vector at the AP. To handle the coupling variables and the non-convexity, the alternating optimization (AO) algorithm is utilized to separate the formulated problem into several subproblems, which are solved by using the semidefinite relaxation (SDR) together with the Dinkelbach methods. Numerical results demonstrate the effectiveness of the proposed AO based CE maximization algorithm. It is shown that the proposed ARIS-UTC outperforms the traditional passive reconfigurable intelligent surface-aided UTC (PRIS-UTC) as well as the non-RIS-aided UTC (nRIS-UTC) without the PRIS and ARIS in terms of the CE.