The matrix multiplication algorithm is an alternative to the [[Floyd-Warshall algorithm]] but is typically suboptimal in complexity. The algorithm leverages a **min plus** ring which modifies typical matrix multiplication by instead of summing taking the minimum and instead of multiplying each term adding each term. Using repeated squaring, the runtime can be as low as $\theta(n^3 \log n)$.