The concept of autonomy is one of the central concepts in distributed computational systems, and in multi-agent systems in particular. With diverse implications in philosophy, social sciences and the theory of computation, autonomy is a rather complicated and somewhat vague notion. Most researchers do not discuss the details of this concept, but rather assume a general, common-sense understanding of autonomy in the context of computational multi-agent systems. In this chapter, we will review the existing definitions and formalisms related to the notion of autonomy. We re-introduce two concepts: relative autonomy and absolute autonomy. We argue that even though the concept of absolute autonomy does not make sense in computational settings, it is useful if treated as an assumed property of computational units. For example, the concept of autonomous agents facilitates more flexible and robust architectures. We adopt and discuss a new formalism based on results from the study of massively parallel multi-agent systems in the context of Evolvable Virtual Machines. We also present the architecture for building such architectures based on our multi-agent system KEA, where we use an extended notion of dynamic and flexibly linking. We augment our work with theoretical results from chemical abstract machine algebra for concurrent and asynchronous information processing systems. We argue that for open distributed systems, entities must be connected by multiple computational dependencies and a system as a whole must be subjected to influence from external sources. However, the exact linkages are not directly known to the computational entities themselves. This provides a useful notion and the necessary means to establish an autonomy in such open distributed systems.
Autonomy In Mas
Autonomy, from the Greek Auto-nomos (auto meaning self, and nomos meaning law), refers to an entity that gives oneself its own laws. In other words, autonomy means self-governance, and freedom from external influence or authority.
Generally, in multi-agent systems there are two basic attempts and formalisations of the concept of autonomy: internal and external views. The internal notion applies the above definition of autonomy to the agent itself, and specifies a set of principles or architectural constraints that are claimed for an autonomous operation of a given agent. The external view takes a different approach. It does not prescribe anything about internals of the agent or agent architecture itself. It is, rather, an assumption that other agents are autonomous in an abstract sense and cannot be controlled/influenced directly. Agent’s behaviour cannot be imposed by any other agent; hence the interactions and agents collaboration must take into account various aspects of the assumed participants’ autonomy. We discuss briefly these two notions below.
The general notion of autonomy is invariant of the usual architectural or behavioural interpretations. In the next subsections, we review proposals for the definition from the internal and the external point of view.