I'm trying to understand the space complexity for IDDFS. This is stated as being O(bd) in many sources, for example: https://ai.stackexchange.com/questions/24478/what-is-the-space-complexity-of-iterative-deepening-search#:~:text=Iterative%20deepening%20search%20(IDS)%20is,is%20O(bd).
Looking at this diagram:
From my understanding, all the blackened nodes from node "B" are still in memory when the algorithm begins looking at node "C" and its children (tree labelled 1) - is this correct? Or once we look at node "C" onwards are they removed from memory?
The fact that that all nodes are blackened out at the end of this Limit (tree labelled 2) makes it seem that node "b" and its children are still in memory.
If they are kept in memory, I understand using Big O notation should show the "worst case scenario" of an algorithm, in this case shouldn't it be O(b^d)? (For example where node "G" is the goal in this example)