Real Time Decision-Making Strategy Based on Process Tracking

This article examines a new data presentation methodology based on a system of rules for decision-making in artificial intelligence systems. First the method is introduced, which consists of many processes and actions and is based on their interaction. The sets of conditions and actions can be changed without changing the operation of the decision-making algorithm. It allows creating an artificial intelligence model that can build an optimal behavior model for various implementations and functionalities. The actions are built based on defined rules, in addition to decision trees, which allow adapting to the current situation. First, the bot searches for a process based on the priority and conditions for starting the process. Then it executes the action with the highest priority, trying to complete it successfully. Data, in the form of conditions, received after the action is performed, is transferred to the process. This information about how successfully the action was performed are entered into a table. Based on the table, a decision tree is built. If the bot completes all the process’s goals, it finishessuccessfully; otherwise, the bot may continue to execute the process or start looking for a new one. To improve the efficiency of decision-making, reinforcement learning algorithms and decision tree construction are used. The results of applying this method are discussed, and a conclusion on the effectiveness of the method and future enhancements are mentioned.This work offers a universal tool for the development of artificial intelligence in the field of computer games.