Bipartite Matching Is in NC
A new paper claims to have proven that the Bipartite Matching problem belongs to the complexity class NC. This breakthrough in theoretical computer science means the problem can be solved in deterministic polylogarithmic time with parallel processors. It resolves a significant challenge in parallel algorithms and derandomization that has been open since the 1980s, making it a compelling read for HN's technically-minded audience.
The Lowdown
A significant advancement in theoretical computer science has been announced, with a new paper asserting that the Bipartite Matching problem is now classified within the complexity class NC. This development, if it stands, marks the resolution of a major open problem in parallel algorithms and derandomization that has persisted since the 1980s.
- The Bipartite Matching problem involves efficiently pairing elements from two distinct sets (e.g., men and women) based on specified preferences.
- Historically, this problem was known to be solvable in polynomial time, and later, it was shown to be in RNC (Randomized NC), meaning it could be solved in polylogarithmic time with parallel processors but required access to random bits.
- The recent paper by Chatterjee, Ghosh, Gurjar, Raj, and Thierauf presents a derandomization of the existing RNC algorithm.
- This derandomization proves that Bipartite Matching can be deterministically solved in polylogarithmic time with polynomially many parallel processors, thus placing it firmly within the NC complexity class. This categorization into NC indicates a much higher degree of parallelizability and efficiency for a fundamental combinatorial problem, potentially paving the way for more robust and deterministic parallel algorithms in various computational applications.