Scalability of Theo

Scalability of Theo

There are two scalability problems in Theo. First, complexity of actions and world states can increase the search space combinatorially; second, as the number of reactive rules grows, performance due to matching decreases (this is an example of the Utility Problem). This is an especially problematic issue for Theo because it caches all of its experiences, many of them not very worthwhile, and leads to a possibly enormous rule-base. The former issue is handled with explanation-based learning, which adds control knowledge to constrain the search time, somewhat alleviating the Utility Problem. The latter question is unresolved in many architectures, including Theo.


Return to the Issues Index for this Architecture.

Return to the top of this architecture.

Go to a Discussion of this issue for multiple architectures.


Current Location: Theo - Issues - Scalability

Go to NEXT page.