Homework Set #1

Agents and Search

 

 

1)      Go to http://www.aaai.org/AITopics/index.html, and familiarize yourself with the Topics page that is maintained by AAAI.  Choose one of the topics from the pull down menu (choose an area in AI that interests you most) and explore the articles, web sites, current news, research, etc. that are linked from the AAAI site. 

 

In 1-2 type-written pages, describe  the area that you chose.  Include the answers to the following questions (when applicable):

a)      Why is this area part of the study of Artificial Intelligence?

b)      What are the challenges of the research in this area for AI scientists?

c)      What are the some of deployed current systems that use AI technology in this area?

 

2)      State Space formulation:

The sliding tile puzzle has some number of black tiles followed by a space followed by some number of white tiles.

B
B
 
W
W
W

      The goal is to have all black tiles to the right of all white tiles. We do not care where the blank spot is.

You can represent the puzzle as a list, where B represents a black tile, W represents a white tile, and O represents the blank.

   The example above would be:    (B B O W W W)
 

      There are two legal moves for this puzzle:

i)        A tile may move to the adjacent location (left or right) if that location is blank. This move has a cost of 1.

ii)       A tile can jump over one or two other tiles into the blank location.  The cost of this move is the number of tiles that were jumped over.

 

 

a)      Give a formulation for this problem in terms of our discussion in class.  Include the definition of:

i)        state

ii)       start state

iii)     goal state

iv)     successor function

 

b)      Show the search tree for the configuration (B O W W) for both depth first and breadth first without taking costs into account, so that you can arrive at a solution to the problem.  Describe for the more complicated pictured example above what you would do to arrive at the shortest path (with costs taken into account).

c)      Extra credit: Implement breadth first search and depth first search so that you can input an initial configuration, and output the path that will lead to the goal state.

 

 

3)      Given the choice of the following search strategies:

 

a)      Breadth-first search

b)      Depth-first search

c)      Best-first search

d)      Hill Climbing

e)      Constraint Propagation

 

Explain which one you would use (and why) for the following problems:

 

a)      8-puzzle

b)      8 queens problem

c)      Math theorem proving

d)      Machine design

e)      Traveling Salesman Problem

f)        Class scheduling problem

 

4)      Problems from AIMA:

 

Chapter 2:  2.5

Chapter 3:  3.1, 3.6, 3.7a,d 3.17

Chapter 4:  4.2, 4.5 (requires reading pages 95-96)