Sequential Decision Making

The field studies decisions made in sequence, where each action can change future states, available information, and later choices. Common settings include resource allocation, healthcare, robotics, online markets, and language-model agents, often with uncertainty, limited budgets, or changing environments. Approaches range from Markov decision processes, reinforcement learning, and bandits to online optimization, planning, causal methods, and optimal stopping. A central challenge is balancing reliable guarantees for safety, fairness, and performance with the flexibility and computational efficiency needed for long horizons, partial observability, and real-world change.