This article advocates the need for a radical rethink of software agent technology by investigating the mechanisms through which knowledge, deliberation, action and control interact to form truly intelligent autonomous agents, be they deliberative, intentional or purely reactive automata/ particles/actors. Using sound logical formal modelling techniques, this work attempts to propose a unified model of multi-agency, which integrates and consolidates various proposed software agent models including; deliberative, cognitive, collectivist and individualist agent perspectives. In particular, the paper focuses on the formal semantics of model-based and emergent regulatory structure of autonomic self-regenerative systems, agents or particles (swarm intelligence). In addition, the paper uses our new scripting language – Neptune and associated development environment, which transforms formal require-ments models of a given agency to executable code.
展开▼
