You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hi,
I've been working with the package specifically for the EightPuzzleDemo methods, including to get statistics to compare the efficiency of the various search algorithms. One thing we noticed was that IDS and DLS do not produce a QueueSize and MaxQueueSize in their respective Instrumentation metrics. I was just wondering if anyone had maybe found a way to get this information? I've tried to adapt the code from QueueSearch, but the problem is that IDS and DLS do not use the Frontier implementation, which appears to be the way the metrics are set for QueueSize.
Any insight anyone might have is sincerely appreciated!
The text was updated successfully, but these errors were encountered:
Hi,
DLS and IDS implement a kind of DFS recursively. They do not use any queue. So, space consumption only depends on the stack size of the VM. Of course, a queue-based DFS variant could be used/implemented instead. In this case, every node should store its depth. Otherwise, the needed depth test would have a negative impact on time complexity. To add such an implementation to the existing library, one could create an own NodeFactory by subclassing the existing one and use an own node class with an additional attribute depth.
Hi,
I've been working with the package specifically for the EightPuzzleDemo methods, including to get statistics to compare the efficiency of the various search algorithms. One thing we noticed was that IDS and DLS do not produce a QueueSize and MaxQueueSize in their respective Instrumentation metrics. I was just wondering if anyone had maybe found a way to get this information? I've tried to adapt the code from QueueSearch, but the problem is that IDS and DLS do not use the Frontier implementation, which appears to be the way the metrics are set for QueueSize.
Any insight anyone might have is sincerely appreciated!
The text was updated successfully, but these errors were encountered: