Memory-Enhanced Evolutionary Strategy for QoS-Driven RIS Control

Abstract

Operating Reconfigurable Intelligent Surfaces (RIS) requires a control strategy capable of coping with a constantly changing propagation environment. The use of a memory-enhanced evolutionary strategy is proposed here to exploit good solutions found at previous times and boost the exploration of the highly-dimensional solution space. The proposed approach does not require channel sensing at the RIS and its potentiality is demonstrated in a small-scale numerical analysis.