Hello, if you have any need, please feel free to consult us, this is my wechat: wx91due
CMPT 417/827 Fall 2024
This exam contains 13 pages (including this cover page) and 6 questions (including a bonus question worth 4 bonus points).
Total of points is 30 plus 4 (bonus).
Grade Table (for teacher use only)
|
Question |
Points |
Score |
| 1 | 4 |
|
| 2 | 5 |
|
| 3 | 5 |
|
| 4 | 9 |
|
| 5 | 7 |
|
| 6 | 0 |
|
|
Total: |
30 |
|
2. (5 points) What are the advantages and disadvantages of a) uniform-cost search and b) pure heuristic search over A* search?Page 2 of 13 Name & ID:
3. (5 points) In MAPF, one approach to optimize a given initial solution using LargeNeighborhood Search (LNS) is to iteratively remove the paths of a selected subset of agents and replan them. Assume that CBS is used for replanning. Experimental obser-vations suggest that replanning the paths of a larger subset of agents generally results in a greater reduction in the total cost of the solution. Based on this, do you think it is always beneficial to replan paths for a very large subset of agents in each iteration?
4. (9 points) We are given a 4-neighbor grid below.
(a) (6 points) In which order do depth-first search and breadth-first search (both with a sensible duplicate-detection strategy) expand the states when searching from s to g? Ties are broken in lexicographic order. That is, A1 is preferred over A2 and B1, and A2 is preferred over B1.
(b) (3 points) In the arrow puzzle, we have a series of arrows pointing up or down, and we are trying to make all the arrows point up with a minimum number of action executions. The only action available to us is to choose a pair of adjacent arrows and flip both of their directions. Using problem relaxation, come up with a good heuristic for this problem.
5. (7 points) Imagine we create a pattern database for tiles 1, 3, 5 for the TopSpin puzzle. We use a multi-dimensional array PDB[ ][ ][ ] (with indices ranging from 1 to 9 for the three dimensions that indicate the positions of tiles 1, 3, 5, respectively) to store the heuristic values. We want to get a heuristic value for getting from a random configuration (Left) to the goal configuration (Right) by looking up the pattern database.
(We get a PDB for 9, 7, 5 for free. Getting tiles 9, 7, 5 from positions 5, 8, 6 to 9, 7, 5 uses the same cost in the pattern space as getting tiles 1, 3, 5 from positions 5, 2, 4 to 1, 3, 5. )
6. (4 points (bonus)) Dummy question.