MASPY: A Python-Based Framework for Developing BDI Multi-agent Systems

Alexandre L. L. Mellado

doi:10.1007/978-3-032-07638-0_18

Return

MASPY: A Python-Based Framework for Developing BDI Multi-agent Systems

2026-01-01

PRE

Alexandre L. L. Mellado *

André Pinz Borges

Gleifer Vaz Alves

DOI：10.1007/978-3-032-07638-0_18

Original

Original request for help

Save

Abstract

En 中文

This paper introduces MASPY, a Python-based framework to develop Belief-Desire-Intention (BDI) multi-agent systems. MASPY offers a declarative way to program agents in Python using an AgentSpeak like language. By implementing the BDI model in Python, the creation of agents that reason and act autonomously is simplified, enabling developers to focus on high-level system behavior. Besides, MASPY benefits from being built on top of a mainstream language like Python, which provides an entire ecosystem and extensive libraries to make the framework accessible and extendable. It presents a thorough multi-agent framework capable of representing major BDI multi-agent features like communication, environment, and plans with a robust BDI reasoning cycle. To validate the framework, we present the context-free grammar of MASPY along with its related concepts, define the agent reasoning cycle to show how a BDI agent works in MASPY, and present a proof of concept through implementing and simulating a well-known multi-agent problem like negotiation. This proof of concept seeks to demonstrate the effectiveness and flexibility of MASPY in building general-purpose multi-agent systems.

Keywords:

Agent Oriented Programming

Multi-Agent Systems

BDI Paradigm

PythonFramework