Lecture 2.3
Logistics
- Fix YAML files by tomorrow at midnight. The YAML file is your group.
- Deadline for project proposal moved to Monday.
- Quiz
- GitHub Classroom
Learning Objectives
- Understand the general behavior of informed search algorithms
- Describe the specific behavior of the Greedy Best-First and A\* Search algorithm
- Be able to define the role of a heuristic
- Name and define criteria for heuristics:
- Admissibility
- Consistency
- (maybe next time) Be able to check whether a heuristic is admissible or consistent
Warm-Up
Q: What’s the difference between an informed and uninformed search algorithm?
A:
Another Warm-Up
Q: What’s a heuristic? Have you used this term before?
A:
Informed Search
(find-file "./slides/informed_search_intro.pdf")
Greedy Best-First Search
(find-file "./slides/gbfs_a_star.pdf")
A-Star Search
(find-file "./slides/gbfs_a_star.pdf")
Heuristics
- A heuristic guides the search process “top-down”
- Often provides “privileged” information (info that can’t be found just by expanding nodes)
Evaluating Heuristics
(find-file "./slides/admissibility_consistency_v2.pdf")
-
Admissibility
-
Consistency
Qualitative Comparison
How do BFS, DFS, Dijkstra, and A* compare in three environments?
What do we think the colors represent?
Search Variants
(find-file "./slides_for_conversion/pdf/a_star_applications.pdf")
Bidirectional Search
Beam Search
Weighted A*
Iterative Deepening Search
Search Summary
<./slides/search_summary.pdf>
(find-file "./slides/search_summary.pdf")