Organization Simulation Based on Normative Knowledge and Role Modeling

An organizational role is defined as a set of functions together with a set of policies (deontic statements). Several roles can be played by the same human or artificial agent. It is also possible that a set of agents is designated to play a given role. In modern organizations it is common to find most agents performing several different simultaneous roles. This may cause ethical and deontic conflicts when the same agent must play conflicting roles. In such cases the agent must use some criterion to solve the internal conflict. In the following we will propose a methodology for approaching this kind of problems, maintaining the essential agents’ autonomy. Organizational agents, humans or machines, must be modeled using a social model. This view is envisioned by the computational semiotics community, where six layers are defined (Stamper 1973) including two broad areas: the information technology platform, which consists of three layers (physical world, empirics and syntax) and the human information functions, which consists also of three layers (semantics, pragmatics and social world). The semantic analysis of business environments can be graphically depicted as an ontology chart, including all agents, roles, resources and their ontological dependencies. An example is shown in figure 1. Additionally to the conceptual knowledge, expressed graphically in the ontology chart, the ontology must also specify behavioral normative knowledge using norms. Norms will be formally defined below, but informally they are seen as social rules that an agent ought to follow. When a clear organizational ontology is not available, where all roles are described and all normative aspects such as authority and power relationships are clear, it might not be possible to understand and predict the agents’ behavior.