Game Artificial Intelligence (AI) is the code in game that makes the computer-controlled opponents (agents) to make smart decisions in game world. There are some AI problems that are essential in games: pathfinding, decision making, generating player’s characteristics, game’s logic management. However these problems are gameplay dependant. The main goal of this study – to generalize AI problems’ solutions for games of many kinds, that is, to make AI solutions gameplay independent. We have achieved this goal by using data-driven design in our solutions. We separated the game logic and the game code levels by using this approach. Such separation gave us an opportunity to manipulate game logic and data freely. We have examined our decision making system and determined that it is flexible and is easy of use.

Game engine is a set of supporting tools and services for game development. It is a component designed for reuse in different games. Therefore it is very important for game engine to be designed properly as for any successfully used reusable component. The main objective in this research is to present flexible and easily extensible architectural solution suitable for the game engine, based on the analysis of today’s game engine context and existing software architecture design. During the analysis we reviewed different types of software architecture and its implementation methods, and defined functional requirements for game engine. As the result we have chosen data-driven architecture design as the most suitable for the engine development and created our game engine prototype. For each game engine module we have provided design solution and its implementation approach referring to chosen architecture.