Using Multi-Agent MicroServices (MAMS) for Agent-Based Modelling

Using Multi-Agent MicroServices (MAMS) for Agent-Based Modelling

연구 분야: Software Development

학회: International Workshop on Engineering Multi-Agent Systems

This paper demonstrates the application of the Multi-Agent MicroServices (MAMS) architectural style to Agent Based Modelling (ABM) through a prototype traffic simulator in which agents model a population of individuals who travel from home to work and vice versa by car. The simulation environment is modelled as a set of linked web resources that are deployed across a number of microservices. The agents, deployed in a separate set of microservices, connect to and interact with their environment using the REpresentational State Transfer (REST). The approach aims to take advantage of various benefits of microservices, such as loose coupling (between the agents and the environments), elasticity (the ability to add and remove environment resources at run-time) and polyglot computing (the ability to use different languages and frameworks for different parts of the application). Finally, the linking of the environment resources leads to the emergence of a simulation wide knowledge graph that can be used by suitably designed agents to supplement their local context with global knowledge of the environment.