For the last few weeks I've been working on the T.D.D.
Mostly doing grunt work like proofreading X_X.
Other than that, I've been documenting the A.I., Josh and I have sorted out pretty much everything you could think of to do with the planning and I've been reading some papers on Heirachical Path-Finding, which actually makes a lot of sense if you read this analogy i found:
"
Consider the problem of traveling by car from Los Angeles, California, to Toronto,Ontario. Specifically, what is the minimum distance to travel by car from 1234 Santa Monica Blvd in Los Angles to 4321 Yonge Street in Toronto? Given a detailed roadmap of North America, showing all roads annotated with driving distances, an A* implementation can compute the optimal (minimum distance) travel route. This might be an expensive computation, given the sheer size of the roadmap. Of course, a human travel planner would never work at such a low level of
detail. They would solve three problems:
1. Travel from 1234 Santa Monica Boulevard to a major highway leading out
of Los Angeles.
2. Plan a route from Los Angeles to Toronto.
3. Travel from the incoming highway in Toronto to 4321 Yonge Street.
The first and third steps would require a detailed roadmap of each city. Step (2) could be done with a high-level map, with roads connecting cities, abstracting away all the detail within the city. In effect, the human travel planner uses abstraction to quickly Find a route from Los Angles to Toronto. However, by treating cities as black boxes, this search is not guaranteed to nd the shortest route. For example, although it may be faster to stay on a highway, for some cities where the highway goes around the city, leaving the highway and going through the city might be a shorter route. Of course, it may not be a faster route (city speeds are slower than highway speeds), but in this example we are trying to minimize travel distance.
"
Other than that the path finding section has been tweaked and little things added. Things are looking good for the hand-in today.
Monday, February 9, 2009
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment