HIHEREI: Human Interaction within Hybrid Environments Regulated through Electronic Institutions Ismel Brito, Isaac Pinyol, Daniel Villatoro, Jordi Sabater-Mir Artificial Intelligence Research Institute (IIIA) Spanish Scientific Research Council (CSIC) Bellatera, Barcelona, Spain {ismel, ipinyol, dvillatoro, jsabater}@iiia.csic.es ABSTRACT In this paper we present an application build on top of electronic institutions in order to create a remote experi- mental platform for human and virtual agents. Our soft- ware objectives are twofold: (1) provide experiment design- ers with a tool to design an experiment where human and virtual agents will interact, and, (2) provide experimental subjects with a friendly interface to interact with virtual agents through an electronic institution. Categories and Subject Descriptors I.2.11 [Distributed Artificial Intelligence]: Multi-agent Systems Keywords Electronic Institutions, Web Interface, Experimental Plat- form 1. INTRODUCTION Disciplines like sociology, psychology, anthropology, econ- omy etc... have based part of their research on observ- ing, monitoring and analyzing individual’s actions in pre- designed and controlled scenarios. Some of them require some interaction among the participants, for instance, in experiments or games where there is a competition for the same products. Nowadays, a lot of these experiments are de- signed using computerized models offering the participants nice interfaces to play with. In these kind of experiments humans sometimes are substituted by artificial agents, both to simplify the complexity associated with human experi- ments with a big number of participants and also to study the reaction of humans in front of these autonomous artifi- cial entities. In these hybrid experiments, humans and au- tonomous agents are put together in the same environment to interact. In order to perform these experiments it is necessary a theoretical and technological framework that can support both the execution of the experiments and the gathering of the results for a subsequent analysis. Given the kind of experiments we are interested to be able to design and Cite as: HIHEREI: Human Interaction within Hybrid Environments Regulated through EI, I. Brito, I. Pinyol, D. Villatoro and J. Sabater-Mir, Proc. of 8th Int. Conf. on Autonomous Agents and Multia- gent Systems (AAMAS 2009), Decker, Sichman, Sierra and Castel- franchi (eds.), May, 10–15, 2009, Budapest, Hungary, pp. XXX-XXX. Copyright c 2008, International Foundation for Autonomous Agents and Multiagent Systems (www.ifaamas.org). All rights reserved. Electronic Institution E-Agent Governor Performative Structure DB Web server DOM SERV ER SIDE CLIENT SIDE JavaScript (Script language) A JAX E-Agent I-Agent Governor Governor SERVLET Staf f Agent Staff Agent Tracker data Institutional data XML (da ta) Figure 1: Human - eI interaction implement, a framework that seems to fit with our needs is that of the electronic institutions [3]. However, the set of tools already provided to specify, de- velop and run electronic institutions do not allow humans to participate remotely either alone or together with au- tonomous agents in the electronic institution. Given that, we have extended the current technological framework to in- corporate it. This work is the implementation of the design presented in [5]. 2. GENERAL ARCHITECTURE Figure 1 shows the main elements of the presented archi- tecture Electronic Institution (eI): The environment that specifies the performative structures and the interac- tion models. Virtual agents (E-Agents): Agents endowed with autonomous behavior that participate in the e-institution through the governors (elements that provide the agent with the interface to interact with the eI and at the same time restrict the possible actions the agent can perform given the current state of the eI). Interface agents (I-Agents): Agents that represent human users in the electronic institution. They also act as web servers and, like E-Agents, connect to the eI by interacting with the governors. This allows a totally distributed approach. Staff agents: Institutional agents in charge of differ- ent aspects related to the well functioning of the eI.