Hybrid Multi-Agent System for Alzheimer Health Care Dante I. Tapia 1 , Javier Bajo 2 , Juan F. de Paz 1 , Juan M. Corchado 1 1 Departamento Informática y Automática, Universidad de Salamanca, Plaza de la Merced s/n 37008, Salamanca, Spain {dantetapia, corchado}@usal.es, fcofds@gmail.com 2 Universidad Pontificia de Salamanca, Compañía 5 37002, Salamanca, Spain jbajope@upsa.es Abstract. This paper presents a dynamic multi-agent system for monitoring Alzheimer patients’ health care in execution time in geriatric residences. The system architecture implements autonomous deliberative case-based agents with reasoning and planning capabilities, with the ability to operate in wireless devices and capable of obtaining information about the environment through RFID technology. ALZ- MAS description and preliminary results of the prototype in a real environment are presented. 1 Introduction Agents and multi-agent systems (MAS) have become increasingly relevant for developing distributed and dynamic open systems. ALZ-MAS (ALZheimer Multi-Agent System) is aimed to improve the efficiency of health care in geriatric residences, increasing the patients’ quality of life. This paper will describe ALZ-MAS, and explain how this distributes and dynamic multi-agent system has been designed and implemented. A case study is then presented, with the system working in a real environment, with initial results and conclusions analyzed. Is also examined the residential health care problem and the possibilities of Radio Frequency Identification (RFID) as a technology for ascertaining patients location in order to generate plans and maximize safety. This paper focuses in the development of deliberative agents using case-based reasoning (CBR) architecture, as a way to implement adaptive systems to improve assistance and health care support for elderly and people with disabilities, in particular with Alzheimer’s. Agents in this context must be able to respond to events, take the initiative according to their goals, communicate with other agents, interact with users, and make use of past experiences to find the best plans to achieve goals. The development of a deliberative agent that incorporates a Case-Base Planning (CBP) reasoning mechanism, derivative from CBR systems [1], specially designed for planning construction is proposed. CBP-BDI facilitates learning and adaptation, and therefore a greater degree of autonomy than that found in pure BDI (Believe, Desire, Intention) architecture [2]. A Proceedings of the International Joint Conference IBERAMIA/SBIA/SBRN 2006 - 1st Hybrid Artificial Intelligence Systems Workshop (HAIS’2006), Ribeir˜ ao Preto, Brazil, October 23–28, 2006. CD-ROM. ISBN 85-87837-11-7 1