A Penalty Alternating Direction Method of Multipliers for Decentralized Composite Optimization | IEEE Conference Publication | IEEE Xplore