Abstract Institutions (also called normative frameworks) provide an effective mechanism to govern agents in open distributed systems. An institution specifies a set of norms, with respect to the achievement of a goal or goals, that regulate agents’ behaviours in terms of permissions, empowerments and obligations. However, in most real circumstances, several institutions probably have to cooperate to govern the same entities simultaneously, which is very likely to give rise to norm conflicts simply if institutions will be designed independently and typically with different goals. In this thesis, we aim: (i) to identify the different ways to combine institutions, (ii) to model those ways formally and computationally by extending an existing model for single institutions, (iii) to detect conflicts in different types of combined institutions automatically, and (iv) to resolve those conflicts via automatic norm revision using an approach based on inductive learning.