0.4.3 Floyd's Algorithm  Shortest Paths
This algorithm is designed to find the leastexpensive paths
between all the vertices in a graph. It does this by operating
on a matrix representing the costs of edges between vertices.
Before we invoke Floyd's algorithm we must build a matrix,
usually in a twodimensional array. If there are n vertices
in our graph, our matrix will be nxn. Each row in the matrix
represents a "starting" vertex in the graph while each column
in the matrix represents an "ending" point in the graph. If
there is an edge between a starting point i and ending point j
in the graph, the cost of this edge is placed in position (i,j)
of the matrix. If we are dealing with an undirected graph in
which all edges are bidirectional, an entry is also made in
position (j,i) of the matrix. If there is no edge directly
linking two vertices, an infinite (or, in practice, very large)
value is placed in the (i,j) position of the matrix to specify
that it is impossible to directly move from i to j.
For example, if we have a graph in which points 1 and 5 are
connected by a bidirectional edge with a cost of 22 units,
we would place the number 22 into positions (1,5) and (5,1)
of our matrix.
Once we have set this matrix up, we use Floyd's algorithm
to compute the shortest distance between all points in the
graph. Floyd's algorithm is given below in C:
int floyds(int *matrix) {
int k, i, j;
for (k = 1; k <= n; k++)
for (i = 1; i <= n; i++)
for (j = 1; j <= n; j++)
if (matrix[i][j] > (matrix[i][k] + matrix[k][j]))
matrix[i,j] = matrix[i][k] + matrix[k][j];
}
When this routine finishes the entries in all positions of
the matrix represent the lowestcost traversal between the
rowvertex and columnvertex.
Floyd's algorithm works by looking for all nondirect paths between
two vertices that have a lessexpensive total cost than the best way
yet found to move between said vertices. If such a path is found, it
becomes the value against which future indirect paths between these
vertices are tested. In the end, each element of the matrix
represents the lowestcost traversal between the vertices it's row and
column represent. Remember that if the graph is directed, so is the
answer in (i,j) of the matrix; moreover, (i,j) may not be equal to
(j,i) in a digraph.
It is clear that Floyd's algorithm takes n^{3} time. In the next
section we will discuss an alternative to Floyd's algorithm called
Dijkstra's algorithm. It is important to note, however, that for
dense graphs (i.e. graphs with many edges) Floyd's algorithm is as
good as or better than Dijkstra's algorithm.
