Evaluate diffrent Fast Downward Heuristics

[lotu]

It is necessarily to evaluate the different search methods and heuristics we can use with fast downward to improve our search. Currently astar(blind()) is used but this should be improved upon. This should involve some metric to determine which technique is best.

[lotu]

The Snaking Ice Trail level in large.dat show cases this problem. It is very simple but our planner can't solve it. As it requires over 20GB of memory to run the astar(blind()) search. If we move the exit closer it is able to solve it, very easily.