MULTI-AGENT GAME THEORY FOR AI-BASED AUCTION SYSTEMS

Abstract

Because of rapid growth of AI-based bidding process systems in digital marketing, power market and e-commerce platforms has an mandatory design of sophisticated and bendable bidding strategies, able of operating in challenging dynamic environments. This thesis finds the application of very common multi-agent game theory as structure for the bidding strategy optimizations in AI driven auctions systems. In multiple agent cases where self-interested and autonomous agents associate in face of strategic doubt and lim ited knowledge, the research increases the fundamentals of tradi tional game theory, such as the Nash equilibrium, dominant strate gies, and last mechanism design. Creation of intricated, flexible bidding strategies that can be functions in demanding, dynamic sit uation has become required due to quick spread the of AI-based auctions systems in the digital advertising, energy markets, and e commerce platforms. The uses of multi-agent game theory as the most common framework for bidding strategies which optimize in AI-driven sale systems is likely to analyzed in this thesis This research which gets extends to the ideas of classical game theory, e.g., Nash equilibrium, dominant strategies and the mecha nism design, to multi-agent systems with an self-interested and au tonomous actors interacting under strategic instability and incom plete knowledge. The rapid expansion of AI-based auction systems in digital advertising, energy markets, and e-commerce platforms has required the development of complex and flexible bidding. In multi-agent environment where self-seeking, autonomous agents interact in face of strategic ambiguity and incomplete knowledge, the research builds upon the concepts of classical game theory, such as Nash equilibrium, dominant strategies, and mechanism design.