Novelty Accommodating Multi-Agent Planning in High Fidelity Simulated Open World

James Chao,Wiktor Piotrowski,Mitch Manzanares,Douglas S. Lange

2023-06-22

Abstract:Autonomous agents acting in real-world environments often need to reason with unknown novelties interfering with their plan execution. Novelty is an unexpected phenomenon that can alter the core characteristics, composition, and dynamics of the environment. Novelty can occur at any time in any sufficiently complex environment without any prior notice or explanation. Previous studies show that novelty has catastrophic impact on agent performance. Intelligent agents reason with an internal model of the world to understand the intricacies of their environment and to successfully execute their plans. The introduction of novelty into the environment usually renders their internal model inaccurate and the generated plans no longer applicable. Novelty is particularly prevalent in the real world where domain-specific and even predicted novelty-specific approaches are used to mitigate the novelty's impact. In this work, we demonstrate that a domain-independent AI agent designed to detect, characterize, and accommodate novelty in smaller-scope physics-based games such as Angry Birds and Cartpole can be adapted to successfully perform and reason with novelty in realistic high-fidelity simulator of the military domain.

Artificial Intelligence,Multiagent Systems

What problem does this paper attempt to address?

The paper primarily explores how multi-agent systems handle the problem of unknown novelty in high-fidelity simulated open worlds. Novelty refers to unexpected phenomena in the environment that may alter its core characteristics, composition, and dynamics, negatively impacting the agents' behavior planning.

The proposed method aims to enable artificial intelligence systems to detect, identify, and adapt to such novelty, particularly in military simulation scenarios. Specifically, the researchers applied a method called Hydra to a high-fidelity military simulator. This method had previously been used in some simple physics-based games (such as Angry Birds and Cartpole). Hydra uses a standardized planning modeling language called PDDL+ to describe the environment and detects inconsistencies between the planning model and the actual environment through consistency checks, thereby identifying novelty.

Once novelty is detected, Hydra initiates a repair mechanism to resolve these inconsistencies by adjusting parameters in the model. The paper describes an example scenario where the effective range of enemy surface-to-air missiles changes while two aircraft are cooperating to destroy an enemy target, violating the friendly forces' original assumptions about the safe distance. By using the Hydra method, the researchers demonstrated how to enable agents to adapt to such novelty to improve mission success rates.

In short, the problem addressed by the paper is how to enable artificial intelligence systems to effectively plan and make decisions when faced with unforeseen novelty, especially in complex military simulation environments.