Research On Intelligent Moving Of CGF Agent

Computer Generated Forces(CGF) are the autonomous and intelligent simulation agents generated by computers in the distributed virtual battlefield environment, which is based on distributed interactive simulation technology. Autonomic decision-making is an important characteristic of CGF, and modeling human behavior is its key research problem. Motion behavior is the basic intelligent behavior of human, so it is necessary to carry out the research of path planning for CGF's motion behavior.This paper primarily study the intelligent path planning method of CGF agent, and with respect to its limitation, two new methods of path-finding are brought forward. One synthetically considers the influence of environment and situation in certain environment, and the other one synthetically considers the distance, the influence of environment-situation and the influence of uncertainty in uncertain environment. Our contributions can be mainly concluded as following:(1) With respect to the output mechanism of CGF agent's behavior, a path planning framework is put forward based on Belief-Desire-Intention model, which concretely describes the inherent state of CGF agent through the belief, desire, intension and their relation, and also describes its outer behavior, which includes perceiving, reasoning, decision-making and learning.(2) The characteristic of battlefield environment is analyzed, and the environment is divided into six kinds of parameters according to its influence on combat action, which simplified environment's complexity. Then, three kinds of parameters, including climate, trafficability and masking, which have correspondingly more influence on the path planning of CGF agent, are chosen to be analyzed. Based on Cellular Automata, the environment model is built, which can quantificationally describe the influence of each environment element on the path planning of CGF agent by using climate influence factor, trafficability influence factor and masking influence factor. Finally, the entire environment influence factor can be got through integrating above three factors, which can quantificationally describe the influence of whole environment on the path planning of CGF agent.(3) The characteristic of battlefield situation is analyzed. Based on Influence Map, the situation model is built with the combat strength of CGF agent as influence value, which can describe the battlefield's situation and provide information for path planning.(4) Through combining the influence of environment and situation with the A* algorithm, a new method of path planning that synthetically considers the influence of environment and situation is presented. Finally, the simulated experiment is done and shows that the path calculated by the new method is more intelligent.(5) Further improve A* algorithm based on Prospect Theory, and present a new method of path planning that synthetically considers the distance, the influence of environment-situation and the influence of uncertainty. Different paths emphasized on distance, emphasized on the influence of environment-situation and emphasized on the risk evaluation of uncertainty can be found by using the presented method. Finally, the simulated experiment is done and shows the rationality of the three different emphasized paths.The presented path planning method is important for the research of CGF behavior modeling and battlefield simulation, and it is a reference for the path planning in the virtual environment. At the same time, this method can also be used on players'path planning in real-time strategy games, and consequently improves games'artificial intelligence.