Sampling a random permutation with restricted positions, or equivalently approximating the permanent of a 0-1 matrix, is a fundamental problem in computer science, with several notable results attained through the years. In this paper, we first improves the running time of the algorithms for a single permutation. We propose a fast approximation algorithm for the permanent of $\gamma$-dense 0-1 matrix, with an expected running time of $\tilde{O}\left(n^{2+(1-\gamma)/(2\gamma - 1)}\right)$. Our result removes the $n^4$ term from the previous best runtime and provides an improvement for $\gamma \geq 0.6$. When $\gamma = o(1)$, our runtime is $\tilde{\Theta}(n^2)$, which is nearly optimal for this problem. The core of our proof is to demonstrate that the Sinkhorn algorithm, a fundamental tool in matrix scaling, can achieve maximum accuracy of $1/\text{poly}(n)$ for dense matrices in $O(\log n)$ iterations. We further introduce a general model called permutations with disjunctive constraints (PDC) for handling multiple constrained permutations. We propose a novel Markov chain-based algorithm for sampling nearly uniform solutions of PDC within a Lov${\'a}$sz Local Lemma (LLL)-like regime by a novel sampling framework called correlated factorization. For uniform PDC formulas, where all constraints are of the same length and all permutations are of equal size, our algorithm runs in nearly linear time with respect to the number of variables.
翻译:暂无翻译