Open multi-agent systems are populated by self-interested agents, developed by different people using different languages and architectures. One way of making these agents conform to some intended collective purpose is to use norms to regulate their behaviour, i.e. : discourage, prevent and mitigate harmful behaviour and facilitate desirable interaction.

In this thesis, we propose a framework ---composed by a formal model, a distributed architecture and languages and algorithms for its implementation--- that extends the notion of an Electronic Institution in order to manage norms in activities, manage the propagation of their consequences among simultaneous concurrent activities, and resolve possible normative conflicts induced by this propagation.