In this lab exercise, you will use a tool in the "CISpace"
Computational Intelligence Lab at UBC in Vancouver
to explore different search algorithms.
You need to go to the CI Space Web site
at
http://www.cs.ubc.ca/labs/lci/CIspace
and invoke the "Search" applet.
Alternatively, you can also download the code,
and run it locally on your machine as a Java application.
The task is to experiment with the different search methods,
and examine the way they work by looking at a number of
examples.
Examine the sample graph specified in the file
CI-Search-CA.txt
(http://www.csc.calpoly.edu/~fkurfess/Courses/CSC-480/F04/Labs/CI-Search-CA.txt).
This graph represents some distances between cities
in California, and the goal is to explore paths between
San Luis Obispo and Lee Vining. You can either copy and paste
the contents of the file into the editing view of the Search applet,
or download the code, run it as a Java application,
and load the example file from there.
For these two tasks, it is advisable to experiment
with different search settings (see "Search Options"
and "Pruning" menus) to eliminate loops and other
efficiency obstacles.
Note: In the Search 4.03 version, it seems to be necessary to press the
"Auto-Search" button repeatedly to get it started.
Evaluation of Search Methods
In the table below, for each search method identify the settings you selected, list the path chosen by the respective search method,
and describe aspects of that method that you find interesting.
Search Method
Settings
Path
Observations
Breadth-First
Depth-First
Lowest-Cost First
Best First
A*
Tracking the Search Process
The table below is useful in keeping track of the search process.
It shows various aspects of the search, like the current path cost, heuristic,
or f-cost, and the nodes that are "visible", but have not been explored yet
(fringe). Please note that not all of these parameters are utilized
by all search methods.
You don't need to use the table for this exercise, but it might be helpful
to be familiar with it since I may use this or a similar format in an exam.