AI Horizon: Intelligence and Reason in Computers
Introduction Home Essays Resources Contact Info
Essays
  Basic CS
General AI
Chess AI
Go AI

Resources
  Source Code
AI Tips and Tricks
Books
Links

Other
  Code Journal
Site Updates

 

This tip submitted by Alex Allain on 2005-11-29.

Anytime Algorithms

One nice feature of some algorithms is that they can be stopped at any time and you will be able to trust that the result obtained at the given point is usable. When you don't know how long your program will be given to run, or you want your user to be able to say "give me the best answer so far", use an anytime algorithm like iterative deepening search rather than an algorithm that must run to completion, like breadth-first search.

This tip submitted by Alex Allain on 2005-09-11.

Do you need results or explanations?

When considering which algorithm to choose for solvinga problem, think not only about the accuracy of each algorithm, but how easy it is to interpret the results! For instance, a neural network will spit back an answer without giving you any explanation at all. On the other hand, an expert system might be able to provide the exact sequence of logical inferences that took place to yield the result.

Sometimes understanding the reason for the answer that the system gives you is as important as getting an answer. This can happen if you are using the algorithm as a tool for analysis--to give you a starting point for further work, rather than to give you the be-all and end-all answer.

© 2001 - 2023
All content is written and published by the people at or affiliated with AI Horizon <aihorizon.com>.

Please see our Privacy Policy