FOCS Conference 2024 Conference Paper
Approximation Algorithms for Noncommutative CSPs
- Eric Culf
- Hamoon Mousavi
- Taro Spirig
Noncommutative constraint satisfaction problems (CSPs) are higher-dimensional operator extensions of classical CSPs. Their approximability remains largely unexplored. A notable example of a noncommutative CSP that is not solvable in polynomial time is NC-Max-3-Cut. We present a 0. 864-approximation algorithm for this problem. Our approach extends to a broader class of both classical and noncommutative CSPs. We introduce three key concepts: approximate isometry, relative distribution, and generalized anticommutation, which may be of independent interest.