A Compositional Framework for Preference-Aware Agents

TL;DR

This paper introduces a compositional framework based on Soft Constraint Automata for modeling preference-aware cyber-physical systems, enabling modular description and composition of components' behaviors and preferences.

Contribution

It presents a novel framework that unifies the description and composition of system components, preferences, and environmental constraints in cyber-physical systems.

Findings

Framework effectively models preference-aware components

Enables modular composition of complex systems

Demonstrated with a patrolling robot example

Abstract

A formal description of a Cyber-Physical system should include a rigorous specification of the computational and physical components involved, as well as their interaction. Such a description, thus, lends itself to a compositional model where every module in the model specifies the behavior of a (computational or physical) component or the interaction between different components. We propose a framework based on Soft Constraint Automata that facilitates the component-wise description of such systems and includes the tools necessary to compose subsystems in a meaningful way, to yield a description of the entire system. Most importantly, Soft Constraint Automata allow the description and composition of components' preferences as well as environmental constraints in a uniform fashion. We illustrate the utility of our framework using a detailed description of a patrolling robot, while highlighting methods of composition as well as possible techniques to employ them.