Online Situation-Determined Agents and their Supervision

Agent supervision is a form of control/customization where a supervisor restricts the behavior of an agent to enforce certain requirements, while leaving the agent as much autonomy as possible. This is done in a setting based on the situation calculus and a variant of the ConGolog programming language that is situation-determined, i.e., the remaining program is a function of the action performed. When an agent may acquire new knowledge about her environment during execution, for example, by sensing, being situation-determined is not sufficient to ensure a unique remaining program. In such cases we need to consider an agent’s online executions, where as she executes the program, at each time point she must make decisions on what to do next based on what her current knowledge is. In this work, we define a notion of online situation-determined agent program that ensures unique remaining configuration in online executions. Moreover, we formalize the definition of the online maximally permissive supervisor, and discuss its properties.