10
Planning, Acting, and Learning
Now that we have explored several techniques for finding paths in graphs, it is time to show how these methods can be used by agents in realistic settings. I first revisit the assumptions made when I first considered graph-search planning methods in Chapter 7 and propose an agent architecture that tolerates these idealized assumptions. Next, I show how some of the search methods can be modified to lessen their time and space requirements—thus making them more usable in the proposed architecture. Finally, I show how heuristic functions and models of actions can be learned.
10.1 The Sense/Plan/Act Cycle
As mentioned in Chapter 7, the efficacy of search-based planning methods depends on several strong assumptions. ...
Get Artificial Intelligence now with the O’Reilly learning platform.
O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.
