Abstract
This chapter discusses how autonomous agents can adopt organizational rules into their reasoning process. Agents in an organization need to coordinate their actions in order to reach the organizational goals. Organizational models specify the desired behaviour in terms of roles, relations, norms, and interactions. We have developed a method to translate norms into event-processing rules of the agents. We propose a modular reasoning model that includes the organizational rules explicitly. Since the agents are autonomous, they will have their own reasoning rules next to the organizational rules. The modular approach allows for meta-reasoning about these rules. We show that this stimulates bottom-up dynamics in the organization.
TopBackground: Agents And Organizations
In our research we use human organizations as inspiration. From this point of view, we consider an organization as a description of roles, relations and interactions to achieve certain coordination. We assume that the agents fulfilling organizational roles are autonomous entities; they have control over their internal state and their behaviour (Jennings, 2000). This implies that the organizational model specifies behavioural guidelines for the agents to assure desired features such as task coordination or information flow. The agents should follow those guidelines, but they are not forced to do so by definition. Researchers in multi-agent systems have introduced the organizational metaphor to achieve coordination between autonomous agents. Organizational models specify coordination mechanisms between agents in abstract concepts, such as roles, relations and norms.
In this section we discuss related work on organizational models. We describe how different approaches allow agents to take up organizational tasks and we describe the consequences for the agents’ autonomy.
Key Terms in this Chapter
Organizational Model: An organizational model describes the objectives and the structure of an organization in terms of roles, norms, relations between roles and interactions between roles. The description does not include descriptions of participants that with fulfil the roles.
Autonomy: Autonomy in the context of agents means that the agent has control over its internal state and over its behaviour. The first implies that the agent controls how external events influence its internal state. The second implies that the agent has a local decision-making process to decide on actions.
Agent: An agent is autonomous entity that actively pursues its goals and is able to interact with others.
Adjustable Autonomy: Adjustable autonomy means that an agent varies the degree of autonomy of its decision-making process. The degree of autonomy depends on how much the decision-making process is influenced by others. Agents can show adjustable autonomy by dynamically dealing with external influences.
Organizational Rules: Organizational rules specify the desired behaviour of the actors in the organization. They are part of the organizational model.
Reasoning Rules: Reasoning rules specify steps in the decision-making process of the agent. There can be several types of reasoning rules with different purposes, for example, goal planning, event processing, or plan revision.
Meta-Reasoning: Meta-reasoning is reasoning about the reasoning process itself. In this chapter, prioritization of reasoning rules is used as an example of meta-reasoning.