Modeling the interaction of virtual agents in distributed artificial intelligence systems

Abstract

Modern distributed artificial intelligence (AI) systems utilize a significant number of virtual agents that must work collaboratively to solve complex tasks. However, existing technologies for organizing their interaction are characterized by certain shortcomings: high computational complexity, simplified operating conditions, poor adaptability to changes, and significant problems in accounting for the diversity of virtual agents and their emotional reactions during decision-making. The purpose of the study is to develop a new approach for organizing virtual agent operations in distributed AI systems that aims to improve their cooperation, coordination efficiency, and adaptability. The methodological foundation of the study was an innovative approach that combined a specialized emotion model containing 100 virtual agents in a two-dimensional space with a complex network of connections between them, with machine learning methods to enhance virtual agent coordination. Computer modeling methods were applied using experiments in the Python programming environment. The research results demonstrate that effective communication methods between virtual agents significantly improve their coordination, and conflicts during task execution are substantially reduced through adaptive mechanisms. The innovative emotion model can achieve high accuracy levels and contribute to the formation of new system behavior that includes sharp changes in collective decision-making processes. It also identifies essential parameters of virtual agent cooperation to ensure stable system operation. The comprehensive approach based on combining rule-based logic with machine learning can effectively improve virtual agent coordination, especially under conditions of their diversity. The AI system demonstrates real capacity for large-scale changes, but is imperfect in reflecting negative emotional states. Such AI system research results are essential for developing autonomous systems, intelligent networks, and collaboration platforms for virtual agents.