Strategies for solving the N-goal problem

The problem of effectively deciding between multiple goals in real-time the "N-goal problem" is discussed in the context of computer game AI. Two related aspects of the problem, the speed of goal selection and the quality of goal selection, are identified, and two related strategies for solving these problems are introduced. The strategy of goal culling based on insistence heuristics reduces the time costs of goal selection. The strategy of automatic goal selection based on machine learning of human gameplay avoids the problem of explicitly formulating rules to maximise decision quality. Each solution strategy is illustrated by concrete examples from the classic arcade game, "Robotron". The aim is to demonstrate that these techniques are practical for actual computer game development.