Each action is assigned an evaluation value that represents the quality of the action. First, the candidate actions are generated from the current situation of the soccer field. In the current implementation, this planning process is only made by a ball kicker (called a kicker hereafter). A lot of ball kicking action plans are generated during the search process and the best action plan is selected based on the evaluation value.
High school simulator 2018 pc download generator#
This method searches for the best sequence of ball kicking actions among several teammate players using a tree-structured candidate action generator and an evaluation function of the candidate actions. In order to model the cooperative behavior among players as action sequence planning, we employed a tree search method for generating and evaluating action sequences performed by multiple players. Then, our recent approaches for acquiring the evaluation function are described. First, the model of cooperative behavior and its planning process are described. This section shows an overview of online search of cooperative behavior implemented in our team. CYRUS2018 from Atomic Energy High School won the second place, and MT2018 from Hefei University won the third place. HELIOS2018 won the championship with 18 wins and 1 loss and 1 draw, scoring 75 goals and conceding 9 goals. Finally, 13 teams participated in the competition. In 2018, up to 16 teams were allowed to pass the qualificaton process. Due to its functions and stability, the 2D soccer simulator is known as a good research and educational tool for multiagent systems and artificial intelligence. Therefore, the performance of a team depends on the decision process of its agents. The 2D soccer simulator does not model the motion of any particular physical robot, but does capture realistic team level strategic interactions. A game consists of 6000 cycles, thus the decision making process of each player is executed about 6000 times. In the game of RoboCup Soccer Simulation 2D League, player agents make a decision at each cycle in real time. Player agents receive a visual sensor message, an aural sensor message and a body sensor message from the simulation server, and can send few types of abstract action command (kick, dash, turn, turn_neck, and so on). Figure 1 shows a screenshot of the 2D soccer simulator. The simulation system Footnote 1 enables two teams of 11 autonomous player agents and an autonomous coach agent to play a game of soccer with realistic rules and game play. The RoboCup Soccer Simulation 2D League is the oldest of the RoboCup competition.
0 kommentar(er)
