There exist a few programming environments dedicated to the creation of multi-agent systems. Some of them are oriented towards communication between distributed systems, and some others are meant for developing simulation models

Agents for scientific simulation

One can distinguish two great fields for scientific simulation. The first is related to the study of co-operation, competition and auto-regulation mechanisms; It attempts to understand the emergence and collective intelligence concepts. The other field deals with modeling and simulation of complex systems in agronomy, ecology, economy and sociology. The other field concerns the modeling and the simulation of complex systems in various areas such as agronomy, ecology, the economy and sociology.

Some examples of multi-agent platforms :


Ascape is a software framework for developing and analyzing agent-based models. In Ascape, agent objects exist within scapes; collections of agents such as arrays and lattices. These scapes are themselves agents, so that typical Ascape models are made up of "collections of collections" of agents. Scapes provide a context for agent interaction and sets of rules that govern agent behavior. Ascape manages graphical views and collection of statistics for scapes and provides mechanisms for controlling and altering parameters for scape models. Ascape pursuits Sugerscape for the field of the economic surveys.


MODULECO is a prototype of multi-agent platform, founded on an original data-processing model, designed to simulate the markets, the social phenomena and the dynamics of the populations.


MadKit is a Java programming environment dedicated to the creation of multi-agent systems. It is oriented towards communication between distributed systems. It is built upon an organizational model based on Agents, Groups and Rôles : an organization is viewed as a framework for activity and interaction through the definition of groups, roles and their relationships. But, by avoiding an agent-oriented viewpoint, an organization is regarded as a structural relationship between a collection of agents. Thus, an organization can be described solely on the basis of its structure, i.e. by the way groups and roles are arranged to form a whole, without being concerned with the way agents actually behave, and multi-agent systems will be analyzed from the ``outside'', as a set of interaction modes. The specific architecture of agents is purposely not addressed.

MadKit provides general agent facilities (lifecycle management, message passing, distribution, ...), and allows high heterogeneity in agent architectures and communication languages, and various customizations


Mobidyc is a tool for building and using agent-based models relating to population dynamics. It can be used by biologists without any computing skills (Vincent Ginot).


StarLogo is a specialized version of the Logo programming language. With traditional versions of Logo, you can create drawings and animations by giving commands to graphic "turtles" on the computer screen. StarLogo extends this idea by allowing you to control thousands of graphic turtles in parallel. In addition, StarLogo makes the turtles' world computationally active: you can write programs for thousands of "patches" that make up the turtles' environment. Turtles and patches can interact with one another -- for example, you can program the turtles to "sniff" around the world, and change their behaviors based on what they sense in the patches below. StarLogo is particularly well-suited for Artificial Life projects.

StarLogo is a programmable modeling environment for exploring the workings of decentralized systems -- systems that are organized without an organizer, coordinated without a coordinator. With StarLogo, you can model many real-life phenomena, such as bird flocks, traffic jams, ant colonies, and market economies.


Swarm is a software package for multi-agent simulation of complex systems, originally developed at the Santa Fe Institute. Swarm is intended to be a useful tool for researchers in a variety of disciplines. The basic architecture of Swarm is the simulation of collections of concurrently interacting agents: "with this architecture, we can implement a large variety of agent based models".

Mobile agents and assistants

Because of the economic repercussions, this category of agents is mainly attached to the commercial field. They allow either an assistance to the user (the synchronization and the management of diaries, e-mail filtering, memory of the user...), or the adaptive search for information (indexing of documents, request in data base, co-operative search), or the commercial activity (E-trade, virtual company, technological survey auction sales...).

Some examples:

The Software Agents group of the MIT Lab Media develops several projects such as NewT (personalization of the news), Amalthaea (for search, control, and the filtering of the sources of information). Letizia (assisting for search on the WEB). Yenta (kind of yellow page where the users can make knowledge and communicate while keeping their protected personal information), etc.

Microsoft Persona Project: A conversational assistant who interacts with the user in natural language. An initial prototype the Peed parrot was produced to answer the requests of the users in the musical field.

Andersen Consulting: LifestyleFinder, InfoFinder (of the agents which learns offers the preferences from the users and them information while being based on their way of life) BargainFinder (a comparison in real time between the prices of the goods offered on various sites proposes).

Aglets (IBM), Odyssey (General Magic)...

For further information...

For further and exhaustive information on the various categories of agents, you can consult the English-speaking links of Alan Wexelblat and Pattie Maes, the UMBC Agentsweb, the or the Agent Society Home Page or the Reynolds pages.


Le Cirad Centre de coopération internationale en recherche agronomique pour le développement
Legal notice © Copyright Cirad 2001-2015