The sum of all width-one matrices

William Q. Erickson,Jan Kretschmann
2023-08-18
Abstract:A nonnegative integer matrix is said to be width-one if its nonzero entries lie along a path consisting of steps to the south and to the east. These matrices are important in optimal transport theory: the northwest corner algorithm, for example, takes supply and demand vectors and outputs a width-one matrix. The problem in this paper is to write down an explicit formula for the sum of all width-one matrices (with given dimensions $n \times n$ and given sum $d$ of the entries). We prove two strikingly different formulas. The first, a ${}_4 F_3$ hypergeometric series with unit argument, is obtained by applying the Robinson-Schensted-Knuth correspondence to the width-one matrices; the second is obtained via Stanley-Reisner theory. Computationally, our two formulas are complementary to each other: the first formula outperforms the second if $d$ is fixed and $n$ increases, while the second outperforms the first if $n$ is fixed and $d$ increases. We also show how our result yields a new non-recursive formula for the mean value of the discrete earth mover's distance (i.e., the solution to the transportation problem), whenever the cost matrix has the Monge property.
Combinatorics
What problem does this paper attempt to address?
The paper primarily addresses the following issues: 1. **Explicit Formula for the Sum of Width-One Matrices**: The authors propose two different methods to calculate the sum \(S(d, n)\) of all width-one matrices (i.e., matrices with non-zero elements distributed along a path from the northwest to the southeast corner) where the size of the matrix is \(n \times n\) and the sum of all elements is \(d\). Width-one matrices are significant in optimal transport theory, especially in relation to the northwest corner algorithm. 2. **Hypergeometric Series Expression**: The first method utilizes the Robinson-Schensted-Knuth (RSK) correspondence to derive an expression involving a fourth-order hypergeometric series \(4F_3\). This formula is suitable for scenarios where \(d\) is fixed and \(n\) increases. 3. **Application of Stanley-Reisner Theory**: The second method transforms the problem into the framework of combinatorial commutative algebra, particularly using the concepts of Stanley-Reisner rings and Stanley decompositions. The formula produced by this method is more efficient for calculations when \(n\) is fixed and \(d\) decreases. 4. **Average Distance of Earth Movers**: Based on the above results, the paper also demonstrates how to directly calculate the average value of the discrete earth mover's distance (i.e., the solution to the transportation problem) when the cost matrix satisfies the Monge property. This application provides a new, non-recursive method to compute this average, which is more flexible and efficient compared to previous methods. In summary, the core contribution of the paper lies in providing effective methods for calculating the sum of specific types of matrices, and these methods can be applied to a key problem in optimal transport theory—the computation of the earth mover's distance.