Modelling Implicit Communication in Multi-Agent Systems with Hybrid Input/Output Automata

TL;DR

This paper extends Hybrid I/O Automata to model implicit communication in multi-agent systems via specialized world variables that represent environmental interactions, allowing for more accurate modeling of agent-environment dynamics.

Contribution

It introduces world variables with spatiotemporal dynamics into Hybrid I/O Automata to model implicit environmental communication in multi-agent systems.

Findings

World variables effectively simulate environmental perturbations.

Parallel composition of world variables involves summation of signals.

The approach is demonstrated through a simple multi-agent system example.

Abstract

We propose an extension of Hybrid I/O Automata (HIOAs) to model agent systems and their implicit communication through perturbation of the environment, like localization of objects or radio signals diffusion and detection. To this end we decided to specialize some variables of the HIOAs whose values are functions both of time and space. We call them world variables. Basically they are treated similarly to the other variables of HIOAs, but they have the function of representing the interaction of each automaton with the surrounding environment, hence they can be output, input or internal variables. Since these special variables have the role of simulating implicit communication, their dynamics are specified both in time and space, because they model the perturbations induced by the agent to the environment, and the perturbations of the environment as perceived by the agent. Parallel composition of world variables is slightly different from parallel composition of the other variables, since their signals are summed. The theory is illustrated through a simple example of agents systems.