Balancing Exploration and Exploitation: Sequential Decision Making in Humans and Machines

Many decisions that people face in their daily lives require balancing exploration, searching for better alternatives, and exploitation, making a selection based on gathered knowledge. Determining when to stop searching is challenging, and people often deviate from optimal strategies. Furthermore, the cognitive processes underlying how people learn to make these decisions are not well understood. In this dissertation, we first establish a behavioral and computational foundation for sequential decision making, demonstrating how a cognitive model of decisions from experience can predict human learning and stopping behavior across a variety of tasks. We then introduce a novel sequential stopping task to shed light on how, when, and why people decide to stop exploring. With this task, we systematically examine some of the factors that may influence stopping behavior and validate the predictions of our cognitive model. Finally, we investigate interventions designed to help people balance exploration and exploitation more optimally. We leverage wisdom of crowds aggregation techniques, where combining the decisions of individuals may lead to superior outcomes as a crowd than as individuals alone, and provide personalized, cognitive AI-driven recommendations for when to stop searching. By integrating behavioral experimentation with computational cognitive modeling, this dissertation not only advances the theoretical understanding of sequential decision making, but also offers actionable insights for improving how people navigate the trade-off between exploration and exploitation.