Abstract
Strategic decision-making in multi-agent interactions inside the Iterated Prisoner’s Dilemma (IPD) is investigated in this work using Monte Carlo simulations. Building on Axelrod’s work, we present a second-generation tournament with stochastic components, including unpredictable game lengths, to evaluate strategy adaptability and resilience. We analyze how uncertainty influences strategic performance by using a comparison between instances with fixed and uncertain times. We identify, using a descriptive approach, methods demonstrating important behavioral differences between deterministic and uncertain settings. The results provide understanding of adaptive learning, response dynamics, and strategic flexibility, so helping to build strong collaborative strategies for artificial intelligence and decision-making systems. Our results highlight the limitations of exclusively deterministic methods and suggest the necessity for adaptive approaches to improve long-term cooperative success.