HW2 key - HW 2 Answer Key Topics: heuristic search,...

Info iconThis preview shows pages 1–3. Sign up to view the full content.

View Full Document Right Arrow Icon
HW 2 Answer Key Topics: heuristic search, heuristics, memory bounded search, local search, adversarial search 1. One of these things is just like the other In class we said that, while there might be hundreds of search algorithms, they're really just variations of a few core ideas. For example, greedy best-first search acts just like depth-first search when the heuristic h(n) = -1 * depth. In the following questions, explain under what circumstances one algorithm behaves just like another one. Short answers are fine. a. When does uniform cost search behave like breadth-first search? All costs are the same (all costs are 0 is a minor error, = 3pts) b. When does A* behave like uniform cost search? When h(n)=0 c. When does local beam search behave like hill climbing? When the number of beams = 1 d. When does stochastic hill climbing (essentially) behave like hill climbing? When the values of all options is the same e. When does simulated annealing behave (from the very start) like first choice hill climbing? Either: 1. When T=0 2. When all options have a positive value 2. Franken-Algorithms Making an algorithm is often like making a recipe - mix one part of algorithm A, one part algorithm B, etc. Many algorithms are just combinations of simpler algorithms or a simple algorithm and a simple idea (e.g., depth-limited depth-first search is depth-first search + a depth limit). For the following algorithms, list which algorithms or ideas (features, etc.) were used to make the following. There should be two parents for each. Short answers are fine. NOTE: We spent an entire class discussing this a. A* uniform cost and greedy best first search
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
b. Stochastic Hill Climbing random walk and hill climbing c. Simulated Annealing random walk and first choice hill climbing d. How are b and c different? simulated annealing takes the first good choice, stochastic hill climbing picks a good choice at random e. IDA* iterative deepening and A* f. Bidirectional search breadth-first search and searching from both start and goal 3. Mechanics a. You’re doing pathfinding on a 300x500 map (150,000 states/nodes). You are moving from position (10,10) to (15,10). The optimal path is a straight line, requires 5 actions and visits 6 nodes. Since this is a grid, at each node, A* expands and calculates f-costs for 3 nodes and selects 1 (exception: looks at 4 nodes at start node, 0 nodes at goal node). A partial trace is below 0: openList=[ (10,10) ] 1. openList=[ (11,10), (10,9), (10,11), (9,10) ] 2. openList=[ (12,10), (11,9), (11,11), (10,9), (10,11), (9,10) ] Assume that when the search is done, the goal node is moved from the open list to the closed list. The question: Once the search is done, 1. How many nodes are in the closed list? 2. Open list? 3. Are there any nodes not in
Background image of page 2
Image of page 3
This is the end of the preview. Sign up to access the rest of the document.

This note was uploaded on 10/23/2011 for the course ENCS ENCS5 taught by Professor Abdelsalam during the Spring '10 term at Birzeit University.

Page1 / 8

HW2 key - HW 2 Answer Key Topics: heuristic search,...

This preview shows document pages 1 - 3. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online