Adversarial Planning

Adversarial planning focuses on developing strategies for autonomous systems operating in environments where an opponent actively tries to disrupt their goals. Current research explores the computational complexity of these problems, investigating algorithms like Stackelberg planning and evaluating the capabilities of large language models in adversarial scenarios, including multi-agent reinforcement learning frameworks and applications like optimal camera placement. This field is significant because it addresses the security and robustness of planning algorithms in real-world applications, such as autonomous vehicles and security systems, where adversarial actions can have serious consequences. The ultimate aim is to create more resilient and secure planning systems capable of handling unexpected or malicious interference.