1 SOFTWARE AGENTS

Agent technologies are considered as fundamental to the realization of next generation computing. A wide range of software engineering approaches have been devised to manage the complexity of distributed systems – among the most significant ones until now being object-orientation. Object-oriented programming provided abstraction level with the encapsulation of data and objects with several abilities including inheritance. Now it is widely recognized that interaction plays crucial role in complex software. In most cases, software architectures contain many dynamically interacting components engaging in complex coordination protocols. Furthermore, with the changing requirements in systems ’ environments (mostly due to optimization of distributed systems for better performance) the need for a level of autonomy emerged. Autonomous components should be able to respond dynamically to rapidly changing circumstances in dynamic and open environments. Software agents take the next step after object-orientation by giving each ‘object ’ the ability to be autonomous, reactive, pro-active and social – now the software components (agents) have their own thread of control, they respond to the changes in the environment, have their internal goals and are communicative. In distributed systems the idea of an agent is often seen as a natural metaphor. Agent-based systems are systems in which the key abstraction used is that of an agent. An agent-based system can contain a single agent, but arguably the greatest potential lies in the application of multi-agent systems [1]. Conceptual shift from notion of single agent to multi-agent systems is essential paradigm to understand how two approaches differentiate. The most important concepts of bot