using a Social Accounting Matrix to analyze the structure of

[ttemel]

I am trying to analyze a Social accounting Matrix using Graph Theoretical concepts. There I use Combinatorica nd GraphUtilities packages. However, I have just dicovered that several of the codes under Combinatorica and GraphUtilities do not work for example, *CostOfPath*.

I would appreciate if you could give me a simple example of using it because inspite of my countless trials it does not work at all.

Regards,
Tugrul

[Jose_Luis_Gomez]

(* Dear Tugrul: I was able to create a small tutorial about the use of CostOfPath *)
(* This is the first time I use Combinatorica, therefore this is for sure Not the most efficient way to do it*)

(* PLEASE SEE THE MATHEMATICA FILE IN: http://homepage.cem.itesm.mx/lgomez/data/combinatorica.nb *)

(* PLEASE SEE THE PDF FILE IN: http://homepage.cem.itesm.mx/lgomez/data/combinatorica.pdf *)

(* Use of CostOfPath and ShortestPath *)
(* By José Luis Gómez-Muñoz
http://homepage.cem.itesm.mx/lgomez/
Mexico *)

(* FIRST EXAMPLE: Equally Weighted Graph *)

(* Load the package: *)
Needs["Combinatorica`"]

(* Create a simple, equally weighted graph, and store it in the variable myfirstgraph: *)
myfirstgraph = CompleteGraph[6]

(* Show the graph, including vertex numbers: *)
ShowGraph[myfirstgraph, VertexNumber -> True]

(* For this simple graph, the cost of each path *)
(* is just the length of the path (a path is a *)
(* list of vertex numbers). The path {1,3,5,6} is *)
(* the edge 1 to 3, then 3 to 5, then 5 to 6, *)
(* so 3 edges, without weights, the path has a cost of 3: *)
CostOfPath[myfirstgraph, {1, 3, 5, 6}]

(* This is the path. It is made of 3 edges, therefore without weights its total cost is 3: *)
ShowGraph[Highlight[myfirstgraph, {{{1, 3}, {3, 5}, {5, 6}}}],
VertexNumber -> True]


(* SECOND EXAMPLE: Set different weights for the edges *)

(* Load the package: *)
Needs["Combinatorica`"]

(* Create a graph with weights. The weights that *)
(* are Not specified will be considered as "one" by commands like CostOfPath: *)
mysecondgraph = SetGraphOptions[
CompleteGraph[6],
{{{1, 3}, EdgeWeight -> 0.04},
{{3, 5}, EdgeWeight -> 0.08},
{{5, 6}, EdgeWeight -> 0.11},
{{1, 2}, EdgeWeight -> 0.03},
{{2, 5}, EdgeWeight -> 0.07}
}]


(* Weight were defined for some of the edges of the *)
(* graph that was called mysecondgraph. Now the CostOfPath is *)
(* calculated by adding those weights: *)
CostOfPath[mysecondgraph, {1, 3, 5, 6}]


(* Cost of another path: *)
CostOfPath[mysecondgraph, {1, 2, 5, 6}]


(* Any weight that was not specified is considered to be "one": *)
CostOfPath[mysecondgraph, {1, 2, 5, 6, 4}]


(* ShortestPath find the path with minimum CostOfPath from one initial vertex to a final vertex: *)
ShortestPath[mysecondgraph, 1, 6]

(* By José Luis Gómez-Muñoz
http://homepage.cem.itesm.mx/lgomez/
Mexico *)

[Jose_Luis_Gomez]

(* Hello again *)

(* I have created another tutorial, clorser to what you need, Tugrul, now showing how to create a weighted-directed-graph from a matrix. The only inconvinience is that disconnected vertices must have a weight (cost) of a large number, say 1000, instead of the usual zero *)

(* PLEASE SEE THE MATHEMATICA FILE IN: http://homepage.cem.itesm.mx/lgomez/data/digraph.nb *)

(* PLEASE SEE THE PDF FILE IN: http://homepage.cem.itesm.mx/lgomez/data/digraph.pdf *)

Needs["Combinatorica`"];

mymatrix = {
{1, 0.10, 0.15, 0.05, 1000},
{0.05, 1, 1000, 1000, 0.05},
{1000, 0.10, 1, 0.15, 1000},
{0.15, 0.05, 1000, 1, 0.20},
{1000, 0.20, 0.05, 0.15, 1}
};

myweights =
Flatten[MapIndexed[
Function[{weight, pos}, {pos, EdgeWeight -> weight}],
mymatrix, {2}], 1];

mygraph = SetGraphOptions[
CompleteGraph[Length[mymatrix], Type -> Directed],
myweights];

CostOfPath[mygraph, {5, 3, 2, 1}]