A corporative memory based on the user profiles Matı ´as Alvarado * , Manuel Romero-Salcedo, Leonid Sheremetov PIMAyC, Instituto Mexicano del Petro ´leo, Eje Central La ´zaro Cardenas 152, San Bartolo Atepehuacan 07730, Mexico Abstract In this paper, each position in the organization has a well-delimited profile defined by, the assigned tasks as well as for the engaged relationships during the process and the organizational domain. Ontologies for organization positions, tasks and application domains are introduced in order to model an Organizational Memory. This Memory is designed/specified through UML/XML diagrams and it is exemplified by a Customer Relationship Management information system. The organizational’s memory reuses the resulting knowledge from experiences abstraction of organization members while laboring at their positions. q 2003 Elsevier Ltd. All rights reserved. Keywords: User profile; Organizational memory; Ontology 1. Introduction The Organizational Memory requires preserving the acquired knowledge of its members. This can be achieved by historical records from member’s experiences while performing their tasks in a position within the organization. In this work, the design of the Organizational Memory starts, based on the pieces of knowledge (contextual information), from organizational positions. It is well known that in spite of that concrete people accomplish concrete tasks, knowledge abstraction, from people experi- ences, is done by working at each organization position. Organizational knowledge and experiences shall be inte- grated from individual knowledge and experiences (Ban ˜ar- es-Alcantara & King, 1997). Thus, beyond who was the specific person, which while working at some position, has got the experience and has abstracted the knowledge, when such a person by any reason leaves his/her position, the obtained knowledge position functioning does not get lost-at least not at all. The main Organizational Memory added value is that members’ knowledge while working at their respective position, could be strategically reused as an organizational advantage. Usually, an organization relies over some structure, integrating people into work units or work teams. The structure is hierarchically defined in accordance of the responsibilities and influence scope of the members. Each member occupies some position profiled on his/her abilities and assigned tasks. Thus, each member accomplishes specific tasks and at the same time it accumulates related experiences. Considering such experiences and in front of new challenges, it is expected the capacity of proposing solutions to those emerging problems dealing with the organization. This work defines ontologies for positions, tasks and application domains for an information system at the Mexican Institute of Petroleum (IMP for Spanish abbrevi- ation) dealing with customer relationships. Prime activities at IMP are applied research and technological deployment, as well as industrial prototyping for oil exploration, production and refining. Ontology definition for some position by the performed tasks, convey as part of the definition, to characterize the abilities and the information used by the position under focus. In the position ontology, the relationships defined are: the task ontology(ies) which is (are) carried out; those related to ontologies from other positions; and those related with ontologies from the application domain. IMP’s Organizational Memory will be modeled from the ontolo- gies and the relationships above mentioned. Previously, definitions are introduced regarding positions, tasks and domains as part of the model which will allow organiz- ational activities and experiences integration. Additionally, a model is defined in order to keep a record of organizational experience both particular and collective. Organizational Memory is modeled using the Unified Modeling Language (UML) (Rumbaugh, Jacobson & Booch, 1999). Positions are expressed as UML actors and roles, whereas their tasks are represented by UML cases of use, at 0957-4174/$ - see front matter q 2003 Elsevier Ltd. All rights reserved. doi:10.1016/S0957-4174(03)00110-6 Expert Systems with Applications 26 (2004) 87–94 www.elsevier.com/locate/eswa * Corresponding author. E-mail addresses: matiasa@imp.mx (M. Alvarado), mromeros@imp. mx (M. Romero-Salcedo), sher@imp.mx (L. Sheremetov).