Game Development Reference
articles should be read whilst wondering what mistakes are in it. The
references section of the articles that have them is a rich hunting ground for
good information, often pointing to papers of higher quality than the
Wikipedia article citing them.
n Your favorite search engines.
This chapter covers a wide range of topics. Every game AI programmer needs to
be fluent in A*. From there, we find that the more sophisticated topics have to be
judged on the basis of whether they will be useful in a project. The important
point is not that machine learning or planning systems are cool, which they are,
but whether they are the right tool for the job. Unlike the topics in prior chapters,
these were not picked for near-universal applicability or ease of understanding.
By the same token, there is no VB code project for a neural network or some
flavor of planner. They are topics that aspiring AI programmers who have made
it this far should strive toward. The code for them would not be in keeping with
the non-threatening nature of the code in this topic, particularly for those novice
AI programmers who come from a background other than hard-core pro-
gramming, such as animators, producers, or even managers who have worked
thus far to gain solid background in game AI.
Answers are in the appendix.
1. What are the two concepts in A* that let us perform the best-first search?
2. What conditions could cause a node to move from the closed list back onto
the open list in A*?
3. When is a machine-learning system easier to implement than a directly
4. What is the major advantage of behavior trees over FSMs?
5. Why does the search run backward in a GOAP system and forward in an