Source-linked AI summary
Movable-Antenna Position Optimization: A Graph-based Approach
Weidong Mei, Xin Wei, Boyu Ning, Zhi Chen, Rui Zhang
TL;DR
The paper addresses received-power maximization for multiple movable transmit antennas, replacing difficult continuous position searches with discrete point selection. It reformulates selection as a fixed-hop shortest path problem with a polynomial-time optimal algorithm and a linear-time sequential alternative. Numerically, the proposed methods outperform fixed-position antenna benchmarks, achieving about 1.1 dB and 2.5 dB higher received SNR than fixed-position antennas with and without selection, respectively, at sufficiently large sampling resolution.
Problem
Optimizing multiple movable-antenna positions is a combinatory problem when the transmit region is discretized, while prior continuous searches face nonlinear channel-position coupling.
Method
The paper models discrete sampling points as graph vertices with CSI-based edge weights, yielding a fixed-hop shortest path formulation solved optimally in polynomial time, plus a linear-time sequential update method.
Results
Approximately 1.1 dB and 2.5 dB higher received SNR than FPA w/ AS and FPA w/o AS, respectively, are achieved when the sampling-point count is sufficiently large.
Takeaways & Limitations
A moderate number of sampling points can achieve near-optimal continuous-search performance, while the sequential update algorithm remains comparable to the optimal graph-based algorithm.
Abstract
from arXiv · showhide
Fluid antennas (FAs) and movable antennas (MAs) have emerged as promising technologies in wireless communications, which offer the flexibility to improve channel conditions by adjusting transmit/receive antenna positions within a spatial region. In this letter, we focus on an MA-enhanced multiple-input single-output (MISO) communication system, aiming to optimize the positions of multiple transmit MAs to maximize the received signal power. Unlike the prior works on continuously searching for the optimal MA positions, we propose to sample the transmit region into discrete points, such that the continuous antenna position optimization problem is transformed to a discrete sampling point selection problem based on the point-wise channel information. However, such a point selection problem is combinatory and challenging to be optimally solved. To tackle this challenge, we ingeniously recast it as an equivalent fixed-hop shortest path problem in graph theory and propose a customized algorithm to solve it optimally in polynomial time. To further reduce the complexity, a linear-time sequential update algorithm is also proposed to obtain a high-quality suboptimal solution. Numerical results demonstrate that the proposed algorithms can yield considerable performance gains over the conventional fixed-position antennas with/without antenna selection.
I. INTRODUCTION
FA and MA technologies enable antenna-position adjustment to address channel variation, but continuous position optimization is nonlinear and difficult. This letter reformulates discrete point selection as a graph problem and provides optimal and lower-complexity algorithms.
- Motivation: FA and MA technologies dynamically adjust antenna positions within a region to improve channel conditions compared with fixed-position antennas.Their movement can circumvent deep-fading positions and reshape the wireless channel.
- Prior work: Channel responses are highly nonlinear functions of antenna positions, motivating iterative methods that search for locally optimal continuous positions.Prior approaches include alternating optimization, successive convex approximation, and gradient descent.
- Proposed formulation: The paper samples the transmit region into discrete ports and uses point-wise channel information to formulate antenna-position selection.Discretization converts continuous movement into a combinatory sampling-point selection problem.
- Proposed method: The discrete selection problem is recast as a fixed-hop shortest path problem and solved optimally by a customized polynomial-time graph algorithm.Graph vertices represent discrete points, while edge weights are defined from channel state information.
- Complexity and performance: A linear-time sequential update algorithm is also proposed to reduce complexity, and both proposed approaches significantly outperform fixed-position antennas with and without antenna selection.The sequential method obtains a high-quality suboptimal solution through sequential point selection.
A. System Model
The system uses multiple movable transmit antennas and one fixed receive antenna in a quasi-static channel. The transmit region is discretized into feasible sampling points, and maximum-ratio transmission evaluates received power for selected positions.
- System setup: The transmitter has N movable antennas and the receiver has one fixed-position antenna in a MISO system.The movable antennas form a linear array of length L, with extension to a two-dimensional transmit array noted.
- System setup: A quasi-static channel allows the movable antennas to reach optimized positions with negligible movement time relative to channel coherence time.The assumption supports position optimization during the channel's slowly varying period.
- Discretization: The transmit array is uniformly sampled into M discrete positions with spacing δs = L/M, and each antenna selects one sampling-point index.The selected position of antenna n is represented by index a_n from the sampling-point set.
- Position constraint: A minimum inter-antenna distance dmin is imposed to avoid mutual coupling, with a_min = dmin/δs assumed to be an integer.The continuous position problem therefore becomes constrained sampling-point selection.
- Beamforming and objective: For selected sampling points, the transmitter applies maximum-ratio transmission, w_t = √P_t h({a_n})/∥h({a_n})∥, to maximize received signal power.The channel vector is assembled from the channels associated with the selected sampling points.
B. Problem Formulation
The paper formulates discrete MA position optimization as a constrained sampling-point selection problem that maximizes received signal power. Its combinatorial complexity distinguishes it from both continuous position optimization and conventional antenna selection.
- The objective is to maximize received signal power by selecting discrete sampling points for multiple transmit MAs.
- Minimum-distance constraints and many sampling points make the problem different from conventional antenna selection.The sampling-point count M is much larger than the number of transmitting antennas selected in conventional antenna selection.
- The discrete problem remains combinatorial because feasible MA positions must satisfy the distance constraints.
- Enumerating all feasible MA positions has exorbitant complexity, limiting practicality for large M and/or N.
- When δs = dmin, the distance constraints disappear and the complexity reduces to that of selecting N antennas from M fixed-position antennas.
III. PROPOSED ALGORITHMS FOR (P1)
The paper addresses the combinatorial optimization problem with two complementary algorithms: an optimal graph-based method and a lower-complexity suboptimal method.
- The proposed solutions are an efficient graph-based optimal algorithm and a lower-complexity sequential update algorithm.
A. Optimal Solution by Graph Optimization
The optimal method recasts MA sampling-point selection as a fixed-hop shortest path problem on a directed weighted graph. Dynamic programming solves this problem optimally in polynomial time.
- The graph uses sampling points as vertices and orders selected MA indices from nearer to farther points.
- An edge connects two sampling points only when their separation satisfies the minimum-distance requirement.
- Edge weights encode channel power gains, so a feasible MA selection corresponds to a path whose weight sum represents the optimization objective.
- Dummy source and destination vertices convert the MA selection into a path from vertex 0 to vertex M + 1.
- The resulting problem seeks an (N + 1)-hop shortest path, unlike a general shortest path problem.
- Because the graph is directed and acyclic, dynamic programming obtains the optimal path in polynomial time with complexity O(NM^2).
B. Suboptimal Solution by Sequential Update
The sequential update algorithm reduces complexity by optimizing MA sampling points one at a time while maintaining minimum-distance feasibility. Its complexity is linear in M for fixed N, but its result can be suboptimal.
- The sequential method updates one initially selected sampling point during each iteration.
- Each update restricts the candidate set to points separated from already updated and not-yet-updated selections by at least amin.
- The selected point maximizes the objective over the feasible candidate set at each iteration.
- The candidate-set size satisfies M − (2amin − 1)(N − 1) ≤ |Ψn| ≤ M, giving complexity between MN − N(N − 1)(2amin − 1) and MN.
- The method may be suboptimal because candidate sets depend on the initial selection and the order of selected sampling points.
IV. NUMERICAL RESULTS
Numerical results evaluate the proposed algorithms against fixed-position antenna benchmarks across sampling resolution, antenna count, array length, and transmit-path count. The proposed methods generally improve received SNR, with gains shaped by sampling resolution, antenna-position flexibility, and channel fading.
- Sampling resolution: Received SNR improves with sampling points, but gains become negligible beyond M ≥48, indicating near-optimal performance with moderate resolution.The suboptimal sequential update algorithm yields comparable performance to the optimal graph-based algorithm.
- Antenna positions: Optimized MA positions avoid some local channel-gain minima and exploit high-gain regions that fixed-position antennas cannot consistently select.Channel-gain differences across positions can exceed 25 dB; FPA antenna selection remains limited by its available locations.
- System dimensions: Received SNR generally increases with the number of MAs and array length, although the proposed algorithms’ gains over benchmarks diminish as antenna count approaches the distance-constrained maximum.The maximum allowable number is N ≤ L/dmin = 12; increasing N from 11 to 12 can decrease proposed-method SNR because flexibility loss outweighs additional antennas.
- Transmit paths: With more transmit paths, received SNR increases for the proposed MA algorithms and FPA w/ AS but remains almost constant for FPA w/o AS.The paper attributes this to stronger small-scale multipath fading and greater channel-gain fluctuation along the transmit array; with one path, all schemes perform identically.
V. CONCLUSION
The letter optimizes MA positions in an MA-enhanced MISO system using discrete transmit-region sampling. It provides polynomial-time optimal and linear-time near-optimal graph-based approaches, with significant gains over conventional FPAs with or without antenna selection.
- The proposed graph-based algorithm solves the discrete MA point-selection problem optimally in polynomial time.
- A sequential update algorithm solves the point-selection problem in linear time with a near-optimal solution.
- The proposed MA algorithms achieve significant performance gains over conventional FPAs with or without antenna selection.The reported gains hold given a moderate number of sampling points for MA positions.
- The graph-based optimization approach is identified for future extension to more general multi-user and MIMO setups.