Joint Optimization for Power-Constrained MIMO Systems: Is Low-Resolution DAC Still Optimal?

This paper explores the joint optimization of precoding, antenna selection, and transmit power within a fixed power budget at a base station (BS). We aim to maximize the sum spectral efficiency in downlink multi-user multipleinput multiple-output systems. The problem is split into two separate sub-problems: joint optimization of antenna selection and precoding direction, and optimization of the transmit power. We then vectorize the precoding matrix and apply approximation techniques to handle the challenges of the problem. We find a superior Lagrangian stationary point to solve the precoding direction sub-problem, and use gradient descent to solve the transmit power sub-problem. Our simulations confirm the algorithm's effectiveness and show that medium-resolution digital-to-analog converters (DACs) with <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$6 \sim 10$</tex> bits can be more powerefficient than the commonly assumed <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$3 \sim 5$</tex> bits when the total power consumption at the BS is considered.