I am reading "search/domains/tsp.lisp" and "route-finding.lisp"
files. The author explains that "h-heuristic" function is a relaxed
version of a minimum spanning tree.
After reading the source code, I found that "h-heurstic" function is
mainly calculated in the function "path-lower-bound". Comparing
with "graph theory book"'s "minimum spanning tree" chapter, I found
that the algorithm is different.
Source codes's algorithm is more simple because it will doesn't
guarantee the final graph is a tree.
I don't know why to use relaxed version, not full version. Does
this "h-heuristic" value is better than full version's "h-heuristic"
How about your idea?