Min-plus matrix multiplication

Min-plus matrix multiplication, also known as distance product, is an operation on matrices.

Given two n × n {\displaystyle n\times n} matrices A = ( a i j ) {\displaystyle A=(a_{ij})} and B = ( b i j ) {\displaystyle B=(b_{ij})} , their distance product C = ( c i j ) = A B {\displaystyle C=(c_{ij})=A\star B} is defined as an n × n {\displaystyle n\times n} matrix such that c i j = min k = 1 n { a i k + b k j } {\displaystyle c_{ij}=\min _{k=1}^{n}\{a_{ik}+b_{kj}\}} . This is standard matrix multiplication for the semi-ring of tropical numbers in the min convention.

This operation is closely related to the shortest path problem. If W {\displaystyle W} is an n × n {\displaystyle n\times n} matrix containing the edge weights of a graph, then W k {\displaystyle W^{k}} gives the distances between vertices using paths of length at most k {\displaystyle k} edges, and W n {\displaystyle W^{n}} is the distance matrix of the graph.

References

  • Uri Zwick. 2002. All pairs shortest paths using bridging sets and rectangular matrix multiplication. J. ACM 49, 3 (May 2002), 289–317.
  • Liam Roditty and Asaf Shapira. 2008. All-Pairs Shortest Paths with a Sublinear Additive Error. ICALP '08, Part I, LNCS 5125, pp. 622–633, 2008.

See also


  • v
  • t
  • e