Position Optimization for Movable Antenna: A New Evolutionary Framework

In this paper, we propose a new position optimization framework based on differential evolution (DE) to optimize the communication performance of movable antenna (MA)-enabled wireless systems. In particular, we address two key issues in the widely used particle swarm optimization (PSO)-based methods, namely, the extremely high computational cost and the vanilla fitness function. Specifically, in contrast to the existing all-in-one individual representation method that encodes all MA positions into a single individual, we propose a new one-in-one representation method that treats each MA’s position as an individual. This effectively reduces both the dimensionality of a single individual and the number of individuals, thereby significantly reducing the computational complexity. Next, to achieve fast algorithm convergence rate, we further devise an efficient adaptive penalty mechanism for the fitness function that imposes a larger penalty/weight on the constraint encountered a stronger violation instead of the traditionally used uniform penalties. Then, we examine the effectiveness of the proposed method in a typical MA-enabled multiuser downlink communication system, wherein a weighted sum-rate (WSR) optimization problem is formulated and efficiently solved by a two-layer method. Finally, numerical results show that the proposed method can achieve over 95% reduction in computational costs compared to the PSO-based method, yet attaining superior performance.