PhD thesis,
A Model for Organizational Interaction: Based on Agents, Founded in Logic
Universiteit Utrecht, The Netherlands, (2004)SIKS Dissertation Series No. 2004-1.
Abstract
The motivation for this research originated from the need to devise
support systems for Knowledge Management (KM) that incorporate the
management of knowledge assets with the facilitation and encouragement
of interaction between people in an open environment. Support systems
for KM cannot be completely designed in advance, and must accommodate
changing needs and tasks and support users with different goals and
requirements. Three main assumptions underlie this research. Firstly,
in many cases, autonomous entities need a social setting in order
to realize their own individual goals. However, and this is our second
assumption, organizations and/or societies have themselves global
goals and requirements, which are not necessarily shared with any
of the participating entities, but must be achieved by the (coordinated)
activity of those individuals. Thirdly, we assume that a process
of negotiation and adjustment is needed between individual and social
requirements and characteristics in order to conjugate individual
autonomy with social requirements and goals. The OperA model presented
here uses the agent paradigm as conceptual design tool. In our opinion,
the concept of agents presents two powerful bases for organizational
interaction because it enables both the reference to any autonomous
entity participating in an interaction (including people), and provides
theoretical models for entities and interaction. OperA provides a
flexible way to represent interaction and role enactment, because
it abstracts from the specific internal representations of the individual
agents, and separates the modeling of organizational requirements
and aims. Contracts are used to link the different models and create
specific instances that reflect the needs and structure of the current
environment and participants. OperA can contribute to the acceptance
and understanding of agent societies in organizations, because it
is based on an organizational, collectivist view of the environment
and uses the agent paradigm to provide a natural way to view and
characterize intelligent systems. The dissertation furthermore presents
an engineering methodology for OperA models. This methodology was
used to apply the OperA framework to different case studies covering
different aspects of interaction in organized environments, ranging
from KM systems to non-monetary markets for the exchange of health
services. This wide range of applications demonstrates the possibilities
of the concept. Furthermore, OperA enables the use of technology
to support interaction and collaboration in KM environments, in ways
that enrich the organization and take in account individual requirements
and motivations. Finally, the dissertation presents a formal theory
for the OperA framework: the language for contract representation,
LCR, based on deontic temporal logic. LCR is a very expressive logic
for describing interaction in multi-agent systems that makes it possible
to describe and verify contracts that specify interaction between
agents. LCR provides a formal framework and integrated semantics
for OperA, at all three levels of society specification (organizational,
social and interaction). The formalism provides a rather realistic
representation of a domain, in the sense that it treats temporal
and communicative aspects and furthermore is able to represent deadlines
and its influence in the behavior of the model.
