The complexity of human artifacts is reaching a level that was merely unthinkable fifty years ago, and is increasing at an astounding pace. This is particularly true for computer-based artifacts, like robots or software systems.<
In order to deal with and possibly manage this ever-growing complexity, two things are basically required: (1) having a vision of where we are heading to, and (2) devising out new metaphors and abstractions powerful enough to model and engineer the systems of new sorts.
Under this perspective, Part I of this tutorial gives a look to some of the most relevant examples of complex systems in our ever day experience, that is, social and biological systems. So, as a first step, some recent results from these fields are examined and possibly related to the evolution of computer-based systems. As a result, notions like society, environment, organisation and information, among the many others, are emphasised, and their role is mapped upon computer-based disciplines and systems.
In Part II, the role of agent-based abstractions and systems is pointed out, first as a point of convergence for disciplines like Artificial Intelligence, Computer Science and Software Engineering, then as powerful metaphors and abstractions for dealing with the modelling and engineering of complex artifacts. In particular, the role of society and environment as first-class entities in the construction of agent systems is stressed, particularly because of their seeming potential in helping scientists, engineers and practitioners to manage the complexity of computer-based systems.

<div>The basics of social issues in agent systems are then sketched in Part III. The notion of coordination is developed as the conceptual framework where agent interactions and social actions can be defined, understood and constructed. The distinction between subjective and objective coordination models and mechanisms is introduced, and their impact on the modelling and engineering of agent systems is outlined.</div>

<div>Finally, Part IV discusses some approaches to the construction of agent societies, by discussing issues like the role of agent infrastructures, the notion of coordination as a service, prescriptiveness and inspectability of coordination laws, and self-adaptation. Some relations with Workflow Management and Activity Theory are drawn in brief, and some examples based on some well-known coordination models are presented.</div>

Tutorial held at the NETTAB 2002 International Workshop “Agents in Bioinformatics”