1: \begin{abstract}

2: This paper considers a distributed optimization problem in the presence of Byzantine agents capable of introducing

3: untrustworthy information into the communication network. A resilient distributed subgradient algorithm is proposed based on graph redundancy and objective redundancy. It is shown that the algorithm causes all non-Byzantine agents' states to asymptotically converge to the same optimal point under appropriate assumptions. A $O(1/\sqrt{t})$ rate of convergence is also established. 

4: \end{abstract}

5: