Multi-agent system for usability improvement of a university administrative system. Jorge Leoncio Guerra Guerra Félix Armando Fermín Pérez Universidad Nacional Mayor de San Marcos Universidad Nacional Mayor de San Marcos Facultad de Ingeniería de Sistemas e Informática Facultad de Ingeniería de Sistemas e Informática Laboratorio de Robótica e Internet de las Cosas Laboratorio de Robótica e Internet de las Cosas jguerrag@unmsm.edu.pe fferminp@unmsm.edu.pe Abstract by the user in a transparent manner, so that the usability of the system is also improved. The implementation of multi-agent systems is one of the specific ways to integrate heterogeneous 2. Proposed implementation information systems, by creating a software agent that performs specific tasks within a field of known The structure of the proposed multi-agent action, as in the case of the travelling salesman system is shown in Figure 1. problem or an agent searching clinical data of a patient. In this paper, the development of a mobile agent is proposed for the implementation of university administrative services across heterogeneous servers, interacting in turn with other intelligent agents in an heterogeneous communication environment. This will generate a single-access interface, which will serve to improve the usability of the system when the client makes an administrative request regardless of the server on which the requested procedure is processed. Keywords: Integration, usability, agents, multi- agent systems, heterogeneous services. Figure 1 – Proposed multi-agent system 1. Introduction Using the GAIA methodology (Zambonelli et The portal of the University Inca Garcilaso de al, 2003) for modeling a system of agents, la Vega (http://www.uigv.edu.pe) is a web agents that will be built on the prototype are application that offers different services defined: available in each of the offices of the a.- Client Agent: captures requirements or institution, which is accessed through service requests from users; interacts with the hyperlinks from Faculties, Research Units, mobile agent to process the request and sends Distance Education, etc. the response to the client system for viewing An integrated system by software agents or joining the client process. (Shiao, 2004); in which, with a single interface, a user can access different services b.- Server Agent: interacts with the server of without having to go through different pages the corresponding office to transfer the request to finish their request; will make possible to from the mobile agent, processes the requested link these services regardless of the Faculty service, and sends the response to the mobile where they are and make the process requested agent. 138 c.- AGMER Agent: mobile agent moving All modules will communicate via web through JADE middleware to interact with services, with interface agents using AXIS 2. clients and servers, has the ability to divide received requests into subtasks performed by 3. Conclusions and future work different servers. The single-interface model for administrative One advantage of using GAIA is that it can be procedure systems is suitable to improve the combined with diagrams from UML as usability of a web portal to reduce the number described in Kang et al (2004), for this reason of interactions. UML diagrams and GAIA models A multi-agent system allows an heterogeneous (Wooldridge et al is used., 2000) will be used implementation of components as well as for its construction. asynchronous communication similar to a Once developed, the model will be request of a university administrative implemented as shown in Figure 2. procedure. JADE is a suitable platform for developing agents due to its ease of use, and adaptability to development methodologies of multi-agent systems. As future work it is proposed to develop indicators to measure implemented services, the frequency of use of the services, and others, that will be collected and stored in the cloud for future analysis using Big Data tools. 4. References Dan Shiao. 2004. Mobile Agent: New Model of Intelligent Distributed Computing. IBM China, October, 2004. Franco Zambonelli, Nicholas Jennings, Michael Wooldridge. 2003. Developing Multiagent Systems: The Gaia Methodology. ACM Transactions on Software Engineering and Methodology. 12(3): 317–370. Figure 2. Components. In the development process is suggested the Miao Kang, Lan Wang, Kenji Taguchi. 2004. Modelling Mobile Agent Applications in UML2.0 use of Eclipse Luna, JDK 8.0_41 and JADE Activity Diagrams, Proceedings of the 6th 4.3.2 for creating agents. In the prototype, it International Conference on Enterprise was initially considered the access to these Information Systems. Porto. Portugal. Vol. IV: basic services for a university student from the 519-522. case study: Michael Wooldridge, Nicholas Jennings, David a. Specialized Library System with Struts 2, Kinny. 2000. The Gaia Methodology for Agent- running on Google App Engine cloud, a PAAS Oriented Analysis and Design. Journal of solution. Autonomous Agents and Multi-Agent Systems, b. PRODI (English Learning Program) vol. 15. Kluwer Academic Publishers, Boston. The registration system with Node.JS, OpenShift, Netherlands. a PAAS solution. 139