Event Tracking for an Intelligent Automated Agent.
Milind Tambe, Paul S. Rosenbloom · 1994
Abstract: In a dynamic, multi-agent environment, an the intelligent agents act as automated pilots automated intelligent agent is often faced with the for the simulated aircraft. These automated possibility that other agents may instigate events that pilots will take part in exercises with human actually hinder or help the achievement of its own goals. To act intelligently in such an environment, an fighter pilots, where they will aid in tactics automated agent needs an event tracking capability to development and training. For effective continually monitor the occurrence of such events and performance in this domain, these automated the temporal relationships among them. This capability pilots must, among other things, continually enables an agent to infer the occurrence of important monitor events in their environment. For unobserved events as well as obtain a better instance, one crucial event is an opponent’s understanding of interaction among events. This paper focuses on event tracking in one complex and dynamic firing a missile at an automated pilot’s aircraft, multi-agent environment: the air-combat simulation threatening its very survival. Yet, the environment. It analyzes the challenges that an automated pilot cannot directly see the missile automated pilot agent must face when tracking events in until it is too late to evade it. Fortunately, the this environment. This analysis reveals some novel constraints on event tracking that arise from complex multi-agent interactions. The paper proposes one solution to address these constraints, and demonstrates it using a simple re-implementation of an existing automated pilot agent. 1.