Online Alternating Direction Method of Multipliers for Online Composite Optimization

TL;DR

The inequalities established for Online-spADMM are used to develop iteration complexity of the average update of the average update of spADMM for solving linearly constrained convex composite optimization problems.

Abstract

In this paper, we investigate regrets of an online semi-proximal alternating direction method of multiplier (Online-spADMM) for solving online linearly constrained convex composite optimization problems. Under mild conditions, we establish ${\rm O}(\sqrt{N})$ objective regret and ${\rm O}(\sqrt{N})$ constraint violation regret at round $N$ when the dual step-length is taken in $(0,(1 +\sqrt{5})/2)$ and penalty parameter $σ$ is taken as $\sqrt{N}$. We explain that the optimal value of parameter $σ$ is of order ${\rm O}(\sqrt{N})$. Like the semi-proximal alternating direction method of multiplier (spADMM), Online-spADMM has the advantage to resolve the potentially non-solvability issue of the subproblems efficiently. We show the usefulness of the obtained results when applied to different types of online optimization problems and verify the theoretical result by numerical experiments}. The inequalities established for Online-spADMM are also used to develop iteration complexity of the average update of spADMM for solving linearly constrained convex composite optimization problems.

Figures (3)

• Figure 1: Comparison of algorithms with respect to the time-average objective regret and time-average constraint violation for OQO.
• Figure 2: Comparison of algorithms with respect to the time-average objective regret and constraint violation for Lasso.
• Figure 3: Comparison of algorithms with respect to the time-average objective regret and constraint violation for TV.

Theorems & Definitions (19)

• Example 1.1
• Example 1.2
• Theorem 2.1
• Proposition 2.1
• Theorem 2.2
• Corollary 2.1
• Lemma 3.1
• Proposition 3.1
• Proposition 3.2
• Theorem 3.1