Has anyone read the book : Artificial Intelligence For Games by Millington, Ian Is it any good ?

Couldn't you go to a book store, find it, and read it to see if it is OK? That's what I do when I am looking for game programming books.

Also, if you can't find it at a bookstore you could look at user reviews online about that book. I found those very helpful at times.

I think this book has even been reviewed here. Since the OP was too lazy to research it himself I'm too lazy to check also. Google.com, google.com :)

Quote: Couldn't you go to a book store, find it, and read it to see if it is OK?

Do you often go to book stores and read the books before you buy them? If you read them in the book store, why buy them afterwards?

Sorry, couldn't resist.

You're not going to retain what you read. Reading through a programming book in the store is a good way of seeing if it's got the information inside that you're looking for. As well as check if it's written in a style you can easily understand.

Plus they're often used for reference so its good to have around.

I enjoyed it and posted a short review at Amazon.co.uk here
In my defence I followed thesame steps as the other reviewer but *not* with the latest source code. Can't be bothered to go back after nearly a year.
There is another review on amazon.com here .
Hope it helps.

Quote: Original post by Los Frijoles
Couldn't you go to a book store, find it, and read it to see if it is OK? That's what I do when I am looking for game programming books.

90% of the game development books I hear about never make it to the local stores, and even those that do are often bought within a week or two. Besides, reading it for yourself is not often as good as getting opinions from others, who may have more experience in the area and who are better judges of the book's quality.

There's a lot of crappy books out there ("introductory") that just show what fuzzy logic is, and neural nets, and other barely usable junk.

Try Multi-Agent systems -- A Modern Approach. After reading it, I began thinking about how RTS AI often needs the player to micro the units in order to focus fire. As an example of good AI, the contract/bid system.

Ever notice how, in the typical RTS, a hoard of units (Scorpion Spam, Zerglings, etc) require micromanagement in order to focus their fire properly? Even then, a player is usually forced by time constraints to bandbox their entire herd, and sequentially target enemies. A waste of firepower in both scenarios. So...What if units could communicate with each other? How could this help? I think best with examples:

Red advances a carrier group onto Blue's planet, with the goal of invading the planet. So we have two lists of combatants, Red and Blue. Red team starts by sorting his enemies list in order of closest to the enemy, and then begins with the conversing:

Red[1] is the lead gun ship, (fighters haven't been released yet), and closest to the nearest enemy, a blue gun ship. Red[1] calculates that it will take five volleys for him to destroy the enemy, and will expose himself to significant return fire doing so. Red[1] examines Red fleet and finds that there is enough firepower to take out that enemy. So Red[1] announces to his fleet that his is now taking bids for a contract on Blue's lead gunship. The contract requires a total of five of his volleys to decimate the enemy in one go.

Next up, Red[2]. Red two first examines the contract board and finds one available (Red[1] did this too, but there weren't any). So Red[2] makes a tentative bid for the contract, offering his firepower as the unit of measure. Then Red[3] gets a turn, and so on through their fleet.

When Red[1] comes up again, he sorts the bids according to firepower (possibly by range as well), and selects the top ones that can make it happen. Then he sends a message to each Red team member who has been selected, making it a binding contract. Then he marks the contract closed on the message board.

When it comes time for a ship to move, it first checks to see if it's in a binding contract, and if so it attacks that target.

Also, if any member of the contract's team loses some firepower (or dies), they flag the contract as 're-evaluate'. the the next member to do evaluations will re-examine if they can still pull it off. If not, that ship announces the contract is null and void, and works on maybe announcing a new contract.

Anyway, it's a good reference book.

Quote: Original post by papamamba
Has anyone read the book: Artificial Intelligence For Games by Millington, Ian. Is it any good ?

It is not perfect, but of all AI books aimed at games programmers it is head and shoulders above the rest.

Original post by Los Frijoles
Couldn't you go to a book store, find it, and read it to see if it is OK? That's what I do when I am looking for game programming books.

quote]

What a stupid answer