TerranFury-
I''m not familiar with A* techniques of any kind (for pathfinding or A*life, or any other variants), so how exactly would the AI know how to calculate pathfinding costs? Also, do you know of any ways for AI to recognize strategic terrain? For example, recognizing high ground, or choke points? Would the terrain have to broadcast information, or would it be easier for the AI to "recognize" elements of terrain and plan accordingly?
Also, the notion that you have of adding and subtracting works if you know where the enemy is.
Military Intelligence @ Strategic, Tactical, & Operational Level
The world has achieved brilliance without wisdom, power without conscience. Ours is a world of nuclear giants and ethical infants. We know more about war than we know about peace, more about killing than we know about living. We have grasped the mystery of the atom and rejected the Sermon on the Mount." - General Omar Bradley
Links to information I've gathered so far off different forums:
Actual Military Info:
http://www.d-n-i.net/second_level/boyd_military.htm
http://155.217.58.58/cgi-bin/atdl.dll/fm/90-10/toc.htm
http://www.geocities.com/Pentagon/6453/doctrinea.html
http://www.sonshi.com/learn.html
http://www.adtdl.army.mil/atdls.htm
http://carlisle-www.army.mil/usassi/welcome.htm
http://www.strategypage.com/prowg/
Tactical AI Info:
http://www.cgf-ai.com
http://www.botepidemic.com/aid/cgf/design.shtml
I've just bought "Game Programming Gems 1", "Game Programming Gems 2", and "AI Game Programming Wisdom", so hopefully they will contain some usefull info on AI and everything else
"I believe; therefore, it is." -True Perception
"The Requested Information Is Unknown Or Classified" -Anonymous
[edited by - Extrarius on April 16, 2002 5:55:16 PM]
Actual Military Info:
http://www.d-n-i.net/second_level/boyd_military.htm
http://155.217.58.58/cgi-bin/atdl.dll/fm/90-10/toc.htm
http://www.geocities.com/Pentagon/6453/doctrinea.html
http://www.sonshi.com/learn.html
http://www.adtdl.army.mil/atdls.htm
http://carlisle-www.army.mil/usassi/welcome.htm
http://www.strategypage.com/prowg/
Tactical AI Info:
http://www.cgf-ai.com
http://www.botepidemic.com/aid/cgf/design.shtml
I've just bought "Game Programming Gems 1", "Game Programming Gems 2", and "AI Game Programming Wisdom", so hopefully they will contain some usefull info on AI and everything else
"I believe; therefore, it is." -True Perception
"The Requested Information Is Unknown Or Classified" -Anonymous
[edited by - Extrarius on April 16, 2002 5:55:16 PM]
"Walk not the trodden path, for it has borne it's burden." -John, Flying Monk
If you decide to go with the "strategic importance map" that I gave as one option, then you can simply factor in high ground when computing this: high ground has a higher strategic importance than low ground. The details of how the strategic importance map is developed will be entirely game-specific, and will probably be defined explicitly by the programmer. Alternatively, the programmer could design the system with variable coefficients that determine the importance of different factors in calculating the influence map, which could then be tweaked by designers to create multiple "personalities," or be "evolved" using a genetic algorithm.
For information on A*, go to http://www.gameai.com/ and check the pathfinding section. Between that and the many posts on that subject already on this board (try the search function for older posts) you should be able to figure everything out. Pathfinding isn't too hard.
Back to generating the strategic importance map: Chokepoint detection could be performed in a number of ways. One that I've heard of is choosing randompoints on the map and using A*, adding to all points within a certain radius of the path. After you do this for many paths, the areas around chokepoints should have more paths passing through them and should therefore have higher values.
How you detect high ground depends very much on how you store your maps. This is easy with heightmap-based terrain. With arbitrary geometry like Quake III, you may use a more general test like "parts of map visible from location."
Finally, if you can confine your influence and strategic importance maps to two dimensions, you will cut down on memory requirements. However, don't rule out the possibility of using them in 3d. Just tweak the resolution. For example, chances are you need greater resolution in the x and z (horizontal/depth) axes than you do in the y (vertical) axis. Exploit things like this to reduce overhead.
Also remember that what I've told you - the influence mapping approach - is just one way of doing things. It happens to be a way that I like, and one that I find conceptually simple, but you may choose another solution.
[edited by - TerranFury on April 16, 2002 6:21:30 PM]
For information on A*, go to http://www.gameai.com/ and check the pathfinding section. Between that and the many posts on that subject already on this board (try the search function for older posts) you should be able to figure everything out. Pathfinding isn't too hard.
Back to generating the strategic importance map: Chokepoint detection could be performed in a number of ways. One that I've heard of is choosing randompoints on the map and using A*, adding to all points within a certain radius of the path. After you do this for many paths, the areas around chokepoints should have more paths passing through them and should therefore have higher values.
How you detect high ground depends very much on how you store your maps. This is easy with heightmap-based terrain. With arbitrary geometry like Quake III, you may use a more general test like "parts of map visible from location."
Finally, if you can confine your influence and strategic importance maps to two dimensions, you will cut down on memory requirements. However, don't rule out the possibility of using them in 3d. Just tweak the resolution. For example, chances are you need greater resolution in the x and z (horizontal/depth) axes than you do in the y (vertical) axis. Exploit things like this to reduce overhead.
Also remember that what I've told you - the influence mapping approach - is just one way of doing things. It happens to be a way that I like, and one that I find conceptually simple, but you may choose another solution.
[edited by - TerranFury on April 16, 2002 6:21:30 PM]
My Website: ai-junkie.com | My Books: 'Programming Game AI by Example' & 'AI Techniques for Game Programming'
The terrain-reasoning / influence-mapping information over at the one link you had, http://www.cgf-ai.com , seems quite similar to what I have been suggesting.
To summarize my ideas and combine them with his, you could use four maps of scalar values:
- Influence
- Strategic Importance
- Traffic
- Past success
Also, one little tidbit I had left out is this: In the influence map, when dealing with large numbers of combatants, lines of zeroes represent fronts. This is just another way of interpreting the data.
To summarize my ideas and combine them with his, you could use four maps of scalar values:
- Influence
- Strategic Importance
- Traffic
- Past success
Also, one little tidbit I had left out is this: In the influence map, when dealing with large numbers of combatants, lines of zeroes represent fronts. This is just another way of interpreting the data.
I recommend the book How To Make War by James F. Dunnigan. It covers basically every aspect of modern warfare, including the intelligence portion that you are looking for. I''m finding it very helpful in creating a realistic wartime simulation for my game (well, except for the sci-fi part).
------------------------------
Omnipotent_Q
"Poor people are crazy. I''m eccentric."
------------------------------
Omnipotent_Q
"Poor people are crazy. I''m eccentric."
------------------------------Omnipotent_Q"Poor people are crazy. I'm eccentric."
This topic is closed to new replies.
Advertisement
Popular Topics
Advertisement