1: \begin{abstract}

2: This work considers {worst-case utility maximization (WCUM)} problem

3: for a downlink wireless system where a multi-antenna base station

4: communicates with multiple single-antenna users. Specifically, we

5: jointly design transmit covariance matrices for each user to

6: robustly maximize the worst-case (i.e., minimum) system utility

7: {function} under channel estimation errors bounded within a

8: spherical region. {This problem has been shown to be NP-hard, and so

9: any algorithms for finding the optimal solution may suffer from

10: prohibitively high complexity.} In view of this, we {seek} an

11: efficient and {more accurate} suboptimal solution for the {WCUM

12: problem}. {A low-complexity iterative WCUM algorithm is proposed for

13: this nonconvex problem by solving two convex problems

14: alternatively.} We also show the convergence of {the} proposed

15: algorithm, and prove its Pareto optimality {to} the {WCUM} problem.

16: {Some simulation results are presented to demonstrate its

17: substantial performance gain and higher computational efficiency

18: over existing algorithms.}

19: \end{abstract}

20: