I need to find shortest paths from from every nodes to every nodes of
a graph.

I think Floyd warshall could be alright but I've got no idea of how
could be the implementation in prolog...

Any idea?? Google didn't helped unfortunately...

Any other ways to solve this problem?

andrea escreveu:

Since this questions is suspiciously similar to someone asking for a
homework:

Do you know it works so you recognize how to describe it?

Do you know enough Prolog so you can represent the graph in Prolog terms?

If you can do what I ask above, you could build a set of predicates to
perform the algorithm.

Yes, a lot. What is exactly your purpose?

On 11 Lug, 20:52, Cesar Rabak yahoo.com.br> wrote:

andrea escreveu:

[snipped]

From the comment above and the implementation below, it seems your
nodes are being represented by integers, right?

On 12 Lug, 23:31, Cesar Rabak yahoo.com.br> wrote:

andrea escreveu:

On 14 Lug, 06:44, Cesar Rabak yahoo.com.br> wrote:

This is the final predicates (for now)

%%%% data la lista delle citta calcolo il percorso piu breve
%%%% 3 casi, 1.lista terminata, 2.piu di due elementi, 3.due elementi
path([],0).
path([H | T],D) :-
T = [H1 | []], shortest_path(H,H1,K,D,T).
path([H | T],D) :-
T = [H1 | T1], shortest_path(H,H1,K,D1,T), path(T,D2), D is D1 + D2.

%%%% TODO necessarie informazioni su dove si passa
%%%% forma finale dei predicati distanza dist(C1,C2,X)
%%%% shortest_path(Src,Dst,K,Dst,Tappe).
shortest_path(X,X,_,0,_). %%%% citta' con se stessa non...

I changed my predicate to one closer to the floyd warshall's
algorithm, now it's like that.

min_dist(_,[],_). %%%% caso base
min_dist(K,[C1 | T1], []) :- cities(C),path_to_idx(C,I),
min_dist(K,T1,I). %%%% ho consumato la sottolista

min_dist(0,Cities,[C2 | T2]) :-
Cities=[C1 | T1], not(dist(C1,C2,Dist)), write_min_dist(C1,C2,1000,
[C2]), min_dist(0,Cities,T2).
min_dist(0,Cities, [C2 | T2]) :-
Cities=[C1 | T1], dist(C1,C2,Dist), write_min_dist(C1,C2,Dist,[C2]),
min_dist(0,Cities,T2).

min_dist(K, Cities, [C2 | T2]) :-
Cities=[C1|T1], K > 0, J is K-1,
sh_path(C1,C2,Diretto,InfraD), sh_path(C1,J,Before,InfraB),
sh_path(J,C2,After,InfraA),
S is Before+After, min(Diretto,S,Diretto) ->
write_min_dist(C1,C2,Diretto,InfraD), min_dist(K,Cities,T2);
append(InfraD,InfraB,Total), write_min_dist(C1,C2,Total,S),
min_dist(K,Cities,T2).

I tried with a very small example

prova(X,Y) :-
dist(X,Y,Z), Z > 2 ->
write('maggiore'); write('minore'),write(Z).

?- prova(0,1).
minore_L136

So in the "else branch" Z is not unified anymore, why?? I dont' find a
suitable explanation..
How can I translate this little piece of code?