Adaptive Artificial Intelligence for Next-Generation Conflict Simulation

Computer wargames and combat simulation software has reached a very high level of comprehensiveness and sophistication in terms of modeling fidelity and the production of accurate results. However, the development of AI (artificial intelligence) for thesesoftware packages has lagged due to the inherent difficulties in interpreting and managing increasingly complex situations.At the same time, however, the increased complexity of the game system makes a reasonably competent and challenging AI more important than ever, principally becasue it makes operating the game much more difficult for human players. Thus, in order for theprogram to be accessible, let alone fun and enjoyable, requires that the AI be capable of assisting the player in a wide range of situations.Competent AI also serves the function of being able to provide users of the software with an effective training and learning tool. By being able to act like a live human opponent, players can see the results or their actions and strategies in a much morerealistic real-world environment, without requiring the time, potential inconvenience or personnel investment required in a multi-human player situation. The primary benefits of the proposed AI improvements will be that existing and future combat simulations and computer wargames will be more accurate in terms of friendly and enemy force actions, will offer more challenges to players, and will also bemore enjoyable to play. All of which will increase the overall use of the software in whatever professional capacity it is used, be it as a training tool, an evaluation platform for weapons system or tactics development, or even as a information device toshowcase the effectiveness of new technologies.The commercial benefits from developing an impoved AI along the lines of this proposal are that computer games will become more enjoyable to play, increasing sales, as well as offering developers the opportunity to easily and quickly adjust computer AImodels for specific situations without having to rewrite significant portions of the AI modeling code. For example, by devising and adjusting a set of standard AI customization values , the differences in the decisions between German, US or Sovietcommanders in World War II can be reflected, resulting in a better game with much less game development time than is currently required.