Package nl.tue.astar

Interface Tail

All Known Subinterfaces:

FastLowerBoundTail

All Known Implementing Classes:

DijkstraTail

public interface Tail

The tail provides the actual estimate for a head. Implementations can choose to compute the estimate each time, but a tail can also be statefull to speed up the estimation process. A default implementation of a Tail is available in the form of nl.tue.astar.impl.DijkstraTail, which always returns 0 as an estimate.

Author:

bfvdonge

Method Summary

Modifier and Type	Method	Description
boolean	canComplete()	returns true if and only if the replay can finish according to this tail, i.e.
int	getEstimatedCosts(Delegate<? extends Head,? extends Tail> d, Head head)	get an estimate of the remaining distance to the target state in the search.
Tail	getNextTail(Delegate<? extends Head,? extends Tail> d, Head newHead, int modelMove, int logMove, int activity)	constructs the new tail based on the two operations from the old head.
<S> Tail	getNextTailFromStorage(Delegate<? extends Head,? extends Tail> d, CompressedStore<S> store, long index, int modelMove, int logMove, int activity)	constructs the new tail based on the two operations from the old state, which is stored in the given store at the given index.

Method Detail

getNextTail

Tail getNextTail(Delegate<? extends Head,? extends Tail> d,
                 Head newHead,
                 int modelMove,
                 int logMove,
                 int activity)

constructs the new tail based on the two operations from the old head. Preferably, the code to compute the new tail is kept as lightweight as possible.

Parameters:

oldHead -

m -

l -

Returns:

getNextTailFromStorage

<S> Tail getNextTailFromStorage(Delegate<? extends Head,? extends Tail> d,
                                CompressedStore<S> store,
                                long index,
                                int modelMove,
                                int logMove,
                                int activity)
                         throws java.io.IOException

constructs the new tail based on the two operations from the old state, which is stored in the given store at the given index. Preferably, the code to compute the new tail is kept as lightweight as possible.

Type Parameters:

S -

Parameters:

store -

index -

m -

l -

Returns:

Throws:

java.io.IOException

getEstimatedCosts

int getEstimatedCosts(Delegate<? extends Head,? extends Tail> d,
                      Head head)

get an estimate of the remaining distance to the target state in the search.

Returns:

canComplete

boolean canComplete()

returns true if and only if the replay can finish according to this tail, i.e. for the case

Returns: