Breadth First Search - Part 2

How to make breadth-first search more efficient using a visited set.

[geterewlij]

Best teacher I've ever seen in my life ever

[azarku]

Thank you so much for the effort you put in your videos, and hopefully, we will see new videos, , I think your videos have the best explanation on youtube

[kiavash]

I am in my last semester of uni! Bless you for this help, 7 min summed up a 50 min lecture

[gabethebabe]

Mr Levine Sir you got the algorithm teaching game on lock!

[BeSharpInCSharp]

Your lecture helped me solve my AI assignment I had hitting my head for few days! thanks much..

[polihayse]

I just stumbled across your channel while trying to learn about time and space complexity of bidirectional search using BFS for my AI class. It's very clear and well organized. Good job.

I was curious if you were planning on doing any time and space complexity breakdowns for any of these search algorithms. I'm going to try to break this BFS algorithm down with my understanding of time/space complexity.

Given a branching factor of b and a solution of depth d, this algorithm has a time efficiency of O(b^(d+1)). If the goal test was applied when the nodes were generated rather than when they were selected for expansion, then the time efficiency would instead be O(b^d). The space efficiency seems to be the same as the time complexity. Also, this version of the BFS technically has the same worst-case scenario as the previous video, but it will obviously be more efficient on average. It would be cool to see an average-case time complexity of the two different algorithms.

If you are going to do a video on time complexities of BFS, it would be interesting to talk about the version of BFS where the goal state is checked during node generation rather than during node expansion and how this affects the overall performance of the algorithm.

[RyanNL]

Thank you for your well explained vids!

[omrifishbein]

Very well explained. Great!
Comments:
There is no mention of the queue that supports the search and the size issue associated with it and how it works.
Why use a new term "visited" and not just stay with "expanded"?

[ask]

thank you sir!
why didn't you check each child node? for example if the goal is one of the childnodes so you stop the search. therefore the visited nodes are S-A Breadth-first search algorithm (Artificial Intelligence A Modern Approach (3rd Edition), p.82 ). thank you again for this video!

[-vp]

thanks man, please post more video related to AI

[basanthassan]

guarantee to be the shortest path, not the less cost path