Quote:

Original post by Kylotan There isn't likely to be a document when all the games are very different! Lots of the standard AI routines will work well for different aspects of such games though.

I know it, But for example we have books for Role-Playing Games or Strategy games. I looking for a refrence like it for tycoons.

Look at reinforcement learning - thats probably nearest to what you need right now.

Personally I think reinforcement learning is not the right way to go. From the few post mortems I've read on tycoon-style games, the problem is basically one of scheduling. You've got finite resource production and you need to balance resource consumption to maximise a multi-variable performance metric. This is the bread and butter of Operations Research. It's also a common problem in manufacturing process optimisation, so you'll find information in systems and process engineering literature as well.

An important aspect of making game AI is that you often need to read a lot of information that doesn't appear to be relevant to games, in order to pull out the useful bits to make your games look, feel and play more realistically. ;)

Quote:

Original post by shahab Quote: Original post by Kylotan There isn't likely to be a document when all the games are very different! Lots of the standard AI routines will work well for different aspects of such games though. I know it, But for example we have books for Role-Playing Games or Strategy games. I looking for a refrence like it for tycoons.

Those genres seem to sell far more, and are less diverse than 'tycoon' games.

I agree but there is not good documents about tycoon games in anywhere but many people talk about it. I think with current resource you must read a lot of diffrent things and use a few of them because no roadmap and no article bank for these kind of game.
It is good if people with good exp. write some articles in gamedev about tycoon games' system, Ai and Algorithms.

Quote:

Original post by Kylotan Quote: Original post by shahab Quote: Original post by Kylotan There isn't likely to be a document when all the games are very different! Lots of the standard AI routines will work well for different aspects of such games though. I know it, But for example we have books for Role-Playing Games or Strategy games. I looking for a refrence like it for tycoons. Those genres seem to sell far more, and are less diverse than 'tycoon' games.

The AI Wisdom books have articles on some of the various decision concepts used in Tycoon games. Some of them are similar to economic decisions that the AI needs to make in RTS games.

That, however, is your problem. You are assuming that there are decision processes that are specific only to tycoon games - and therefore you continue to ask for those. Really, there are not a lot of AI techniques that are endemic to tycoon games - largely due to what has been told to you repeatedly in this thread... there's too many variations of "tycoon" games. Let's face it... Railroad Tycoon is different than School Tycoon which is different than Skate Park Tycoon which differs from my game, "Airline Traffic Manager". (Airline Tycoon was taken by a pretty pathetic game). Even comparing two transportation sims... mine and Railroad Tycoon... we have very different issues to solve which gives us different AI.

I think your approach should be to learn general AI techniques. Those are your tools in the toolbox. Also, learn how to break a problem down into component parts and then glue it back together. Once you have the tools and the thought process down, you can address almost any issue that should arise - be it in a FPS, RTS, TBS, or a "Tycoon" game.

Quote:

Original post by InnocuousFox The AI Wisdom books have articles on some of the various decision concepts used in Tycoon games. Some of them are similar to economic decisions that the AI needs to make in RTS games. That, however, is your problem. You are assuming that there are decision processes that are specific only to tycoon games - and therefore you continue to ask for those. Really, there are not a lot of AI techniques that are endemic to tycoon games - largely due to what has been told to you repeatedly in this thread... there's too many variations of "tycoon" games. Let's face it... Railroad Tycoon is different than School Tycoon which is different than Skate Park Tycoon which differs from my game, "Airline Traffic Manager". (Airline Tycoon was taken by a pretty pathetic game). Even comparing two transportation sims... mine and Railroad Tycoon... we have very different issues to solve which gives us different AI. I think your approach should be to learn general AI techniques. Those are your tools in the toolbox. Also, learn how to break a problem down into component parts and then glue it back together. Once you have the tools and the thought process down, you can address almost any issue that should arise - be it in a FPS, RTS, TBS, or a "Tycoon" game.

thank you InnocuousFox

best wishes;)