Multi-agent system generating method based on graphical expression

Abstract

The invention relates to a multi-agent system generating method based on graphical expression. The method comprises the following steps of: firstly editing and building a self-defined multi-agent system in a graphical mode; then performing modeling on graphical elements of the multi-agent system; describing the graphical multi-agent system through an expression model; next converting the expression model of the multi-agent system into a middle model described by UML (Unified Modeling Language); and finally converting the middle model into underlying source code, wherein in the code conversion process, corresponding code blocks can be automatically generated according to the mutual relationships between parameters and elements set in the graphical mode, and interaction logics generated by interaction between tasks and agents and between the tasks through service calling are encapsulated. The multi-agent system generating method based on graphical expression can effectively help a multi-agent system developer understand and use the multi-agent system; the development workload is reduced; the complexity degree of the interaction logics is reduced; and the development efficiency is improved.

Description

technical field [0001] The invention relates to a multi-agent system development technology, in particular to a graphical representation-based multi-agent system generation method. Background technique [0002] Multi-Agent System (MAS for short) is an important branch of distributed artificial intelligence. By organizing multiple agents together, communicating and coordinating with each other, they can jointly achieve some large and complex goals or complete specific tasks. task. The multi-agent system has the characteristics of autonomy, distribution, reactivity, and initiative, and has certain organizational, learning, and reasoning abilities. Moreover, different agents can be heterogeneous and distributed. No matter what environment and platform a single agent runs on, as long as the agents communicate and interact according to the agreed specifications, they can form a multiple intelligence. body system. Because multi-agents not only have the ability to solve problems...

AI Technical Summary

Problems solved by technology

However, there are two problems in the current graphical development method. First, the representation of the graphical multi-agent system model involves too many details of the internal mechanism of the multi-agent system, such as the negotiation mechanism, etc., and the degree of abstraction is not high. It is beneficial for users to understand and use, and requires users to master a large amount of computer expertise about multi-agent systems; Converted to C++ code, and many software in industrial fields, such as the development of software related to intelligent distribution network, use C++ as the mainstream development language, which greatly restricts the application range of graphical development methods

Images