N.T. Nguyen et al. (Eds.): KES-AMSTA 2007, LNAI 4496, pp. 92–101, 2007. © Springer-Verlag Berlin Heidelberg 2007 A Methodology to Specify Multiagent Systems Aguilar Jose 1 , Cerrada Mariela 1 , and Hidrobo Francisco 2 1 CEMISID (Centro de Microcomputacion y Sistemas Distribuidos), Facultad de Ingeniería, Universidad de los Andes, Mérida 5101, Venezuela {aguilar, cerradam}@ula.ve 2 SUMA, Facultad de Ciencias, Universidad de los Andes, Mérida 5101, Venezuela hidrobo@ula.ve Abstract. In this paper we present a methodology to specify Multiagent Sys- tems, called MASINA. MASINA is based on MAS-CommonKADS; we use the models presented in this methodology to propose some extensions, modifica- tions and substitutions allowing to describe the intelligent characteristics of an agent or group of agents, to use intelligent techniques for the accomplishment of tasks (e.g. artificial neural networks), and to specify emergent coordination approaches, among others. 1 Introduction The investigation of methodologies for analysis and design of Multiagent Systems (MAS) is still in embryonic state. The existing methodologies for developing MAS are not exactly new; generally they are extensions from object-oriented methodolo- gies or knowledge engineering methodologies, given their close relationship. The existing ones have important weaknesses and disadvantages which makes them un- able for usage in complexly designed environments. In this article we propose a methodology for the specification of Multiagent Systems, named as MASINA (Mul- tiAgent Systems in Automation). The term of Automation is due that this methodol- ogy was originally developed to solve automation problems using MAS [1, 2]. This methodology is based on MAS-CommonKADS [3], proposing some extensions, modifications and substitutions of the models that have already been defined in MAS- CommonKADS. MASINA depicts the fundamental elements in the MAS area, such as representing the notion of intelligence of an agent (intelligent agents modeling), in a collective level (swarm intelligence), the coordination mechanisms between agents (emerging planning, conflict resolution, etc.), the direct or indirect communication, along with other things. In addition, MASINA allows characterizing other aspects such as the use of intelligent techniques for carrying out tasks, the use of reference models for specifying agents and the definition of conversations between agents as well as the speech acts (interactions) within these conversations. 2 Multiagent Systems Specification Methodologies There are three main groups of propositions of designing methodologies based on agents, the first one is based on object-oriented methodologies, the second one is based on