Lab: Online Search

CSC 261 - Artificial Intelligence - Weinman

Summary:: You will play the role of a blind agent, running an online depth-first-search algorithm

Preparation

Open AIMA to page 150 and review the ONLINE-DFS-AGENT algorithm in Figure 4.21.
Launch the exploration program:
1. In your terminal:
  
  $ drracket /home/weinman/courses/CSC261/code/online/run-online-dfs.scm &
2. Click "Run".

Exercise

You are an agent traversing an unknown environment using the ONLINE-DFS-AGENT algorithm. At each state you will be asked what action to take according to the algorithm. If you answer incorrectly, you will be prompted until you answer correctly. Then you'll have to answer why that particular action was taken.

To do backtracking correctly, you must keep a diagram of the state space you have seen so far (that is, a state-action transition diagram like AIMA Figure 3.3 on p. 70). The program is doing this behind the scenes for you. When we're done, one or two groups will draw their state spaces on the board.

Reflection

Be prepared to discuss the following questions

What role does the unbacktracked queue play?
How would you characterize the behavior? Does it seem efficient? Are there alternatives?
How hard is this particular problem (e.g. state space, actions, etc.)?
- What if we had more states?
- What if we had a continuous environment?

This work is licensed under a Creative Commons Attribution-Noncommercial-Share Alike 4.0 International License.