Book on AI
The book is still very much in the planning stage. Release date is expected to be in the first half of next year, although the book won’t be rushed; it will only be released when I believe it:
For more information see:
The goal of the book will be to teach existing (we won’t be touching on basic programming techniques/methodology) programmers a range of AI techniques by bringing a small adventuring village and its surrounding area to life. At first the village is empty and abandoned but as the book progresses, each section of the village is brought to life; from the small town itself (complete with a blacksmith, shop keepers, hunters, and various other NPC’s), to its surrounding areas (including a farm, the obligatory dungeon, and the wilds with free roaming predators and prey). While the contents are subject to change (which they seem to be doing weekly at this early stage), the core topics that I aim to cover include: Procedural Dungeon Generation: create an interesting environment for your player to explore. Also includes clever methods of populating these environments (not just selecting a random floor tile!). Implementing a modular Finite State Machine (FSM) for enemy AI. With detailed instructions on how this can be implemented in any future game. Goal Oriented Action Planning for NPC AI. We’ll give the towns people their own personal goals and let them work out what they want to do at any given moment. Path finding: it doesn’t matter how clever we make the towns people and dungeon dwellers, if they keep walking into the wall they won’t seem all too smart. Procedural behaviours such as obstacle avoidance, and group behaviours (e.g. flocking) for animals in the game: Let’s create a small farm for the village and give the animals a way to move through the environment. Lastly we’ll populate the wilds with a predator and prey ecosystem: These animals will be controlled by an Artificial Neural Network (ANN) and evolved using a Genetic Algorithm (GA) to simulate natural evolution. The ecosystem will evolve as the game is played and the predators and prey will (hopefully) become more efficient. I say hopefully because things can become a bit difficult to predict when we enter the domain of neural networks and evolutionary computation.
Robert Wells
Founder of Gandhi Games. - Programmer