Page 323 - Kaleidoscope Academic Conference Proceedings 2024

MISO-BASED MEC SYSTEM WITH IRS:
JOINT COMPUTATION OFFLOADING AND BEAMFORMING OPTIMIZATION

Zhiguo Xu; Shucheng Wang; Zhongming Ji; Jun Xu; Dongxu Wang; Shengju Zhang; Lingkun Meng
China Mobile Communications Group Co., Ltd, China

ABSTRACT achieving reconfigurable wireless propagation environment
[6]. Compared to traditional active relays which regenerate
Intelligent reflecting surface (IRS) has appeared as an and retransmit signals, IRS only passively reflect incident
innovative green and economical technique to markedly signals with low energy consumption. Besides, IRS operates
enhance the spectrum and energy efficiency in mobile in full duplex (FD) mode with suppressing interference [7].
edge computing (MEC) systems. However, limited Because of these advantages, IRS is capable of improving
by single antenna setting, the potential of IRS-aided the communication rates, and mitigating latency of MEC
MEC systems has not been fully exerted. To further systems. Thus, the combination of IRS and MEC can
improve the system computational performance, the paper reinforce their respective advantages and further enhance
introduces multiple-input single-output (MISO) technology, the potential of MEC systems. Also, the new paradigm
and proposes a MISO-based MEC system with IRS. In will be applied in diverse scenarios of B5G/6G wireless
this system, IRS is utilized to support the communication communication systems. [8] minimize the computational
between users and a multi-antenna access point integrated latency of an IRS assisted MEC system by jointly optimize
with edge servers. Also, we adopt non-orthogonal multiple computing and communications in different scenarios. [9]
access (NOMA) strategy for information transmission. The formulated an energy minimization problem of IRS-aided
computation rate is maximized to optimally design receiver MEC systems to optimally design the local CPU frequency,
beamforming, CPU frequency, transmit power, and IRS receiver beamforming, offloading schedules and IRS phase
phase shifts. This formulated problem is non-convex and shifts. [10] jointly optimized time assignment, transmission
challenging to solve due to multi-variable coupling. To tackle power, and IRS phase shifts to maximize total computational
it, we exploit alternating optimization manner to address the bits of IRS-aided MEC system, and presented the introduction
four decoupled subproblems until convergence. Simulation of IRS can achieve significant advantages.
results highlight the superior computation rate performance
Meanwhile, the advent of non-orthogonal multiple access
of our proposed IRS-aided MISO-based MEC system.
(NOMA) technique alleviate the problem of limited spectrum
Keywords - Computation rate, IRS, MISO, MEC, NOMA resources in IRS-assisted MEC systems, which can improve
spectral efficiency, and support large scale connectivity [11].
1. INTRODUCTION Therefore, the involvement of NOMA can complement and
synergize with the advantages of MEC, effectively ascending
With the popularity of mobile devices, the demands for high the system transmission rate and computational efficiency.
information transmission rate and QoS are increasing [1]. [12] introduced IRS to assist task offloading for two users
Also, the continuous emergence of computation-intensive in MEC system, in which users transmit data in NOMA and
and latency-sensitive applications (e.g., image/biometric time division multiple access (TDMA) strategies, and the
recognition, and augmented/virtual reality) has brought new IRS phase shifts and offloading scheduling was optimized
challenges to B5G/6G wireless systems [2]. Aiming at coping to minimized user delay. [13] considered an IRS-assisted
with these challenges, mobile edge computing (MEC), as an MEC system where users communicated with access point
up-and-coming technique, has been proposed to significantly (AP) through the NOMA protocol to effectively utilize
improve the computational capability for mobile devices with spectrum resources. In particular, [13] jointly optimized
limited computing resources [3]. In MEC systems, small-size IRS phase shifts, decoding order, transmission data, power
and power-limited devices offload their intensive tasks to and time to minimize the sum energy consumption of all
nearby powerful edge servers, and these tasks can be executed users, and demostrated the proposed scheme achieved marked
remotely and downloaded, thus efficiently reducing latency performance gains compared to the IRS-aided MEC system
and expanding computational capacity [4]. with TDMA protocol. However, the abovementioned works
Additionally, intelligent reflecting surface (IRS), an all adopts single antenna technology, which still cannot
innovative green and economical technology that has fully explore the potential of IRS-aided MEC systems.
emerged lately, can markedly extend communication Compared with single antenna technique, multiple-input
coverage and enhance energy-efficiency [5]. In general, single-output(MISO) technology can efficiently strengthen
IRS consists of massive electromagnetic units, which can communication quality, and increase the capacity and spectral
independently change the amplitude and/or phase, thus efficiency [14]. Besides, computational rate is also the

978-92-61-39091-4/CFP2268P @ITU 2024 – 279 – Kaleidoscope