Towards Scalable Semidefinite Programming: Optimal Metric ADMM with A Worst-case Performance Guarantee

Yifan Ran,Stefan Vlaski,Wei Dai

2024-03-18

Abstract:Despite the numerous uses of semidefinite programming (SDP) and its universal solvability via interior point methods (IPMs), it is rarely applied to practical large-scale problems. This mainly owes to the computational cost of IPMs that increases in a bad exponential way with the data size. While first-order algorithms such as ADMM can alleviate this issue, but the scalability improvement appears far not enough. In this work, we aim to achieve extra acceleration for ADMM by appealing to a non-Euclidean metric space, while maintaining everything in closed-form expressions. The efficiency gain comes from the extra degrees of freedom of a variable metric compared to a scalar step-size, which allows us to capture some additional ill-conditioning structures.

Optimization and Control

What problem does this paper attempt to address?

This paper attempts to address the issue of improving the scalability and efficiency of the Alternating Direction Method of Multipliers (ADMM) in Semidefinite Programming (SDP). Specifically:

1. **Existing Problem**: Although Semidefinite Programming has wide applications in many fields and can be solved using Interior Point Methods (IPMs), the computational cost of IPMs is very high when dealing with large-scale data. While first-order algorithms like ADMM can alleviate this issue, their scalability is still insufficient in some cases.

2. **Research Objective**: This paper aims to further accelerate ADMM by introducing non-Euclidean metric spaces while keeping all iterative formulas in closed form. This additional acceleration comes from the fact that variable metrics have more degrees of freedom compared to scalar step sizes, allowing the algorithm to capture some additional pathological structures.

3. **Application Background**: The paper considers Quadratically Constrained Quadratic Programming (QCQP), a problem that naturally appears in the form of SDP. Through numerical experiments on this type of problem, it is observed that scalability is significantly improved, especially when the data conditions become worse, making this advantage more apparent.

In summary, this paper is primarily dedicated to enhancing the performance of the ADMM algorithm on large-scale, ill-conditioned data by improving its design.