A framework for institutions governing institutions

Conference item

Norms guide multi-agent systems away from being potentially anarchic towards a coordinated and collaborative society. Institutions provide an explicit, external representation of norms as well as the means to detect violations and other conditions. Each institution can be crafted individually to capture their designers’ goals, but this creates a challenge at higher levels of authority in guiding the institutional design to be coordinated with other institutions and not imposing unacceptable limits on agents’ rights. We propose to use institutions to govern and to revise institutions, following a principle widely encountered in the social world, where treaties, primary legislation, framework agreements and subsidiarity establish a regulatory space by defining norms on the form of a body of regulation. We set out a formal and computational framework, building on the InstAL model and implementation, to construct tiers of institutions, where the norms at each tier are governed by those at the tier above. Thus, agents’ behaviour is governed and monitored by a tier-1 institution, whose norms are governed and monitored by a tier-2 institution, etc.. This allows us to check the compliance of an institution with the tier above. Compliance failure generates the necessary negative examples for automatic norm-revision.

Authors: King, T; Li, T; De Vos, M; Dignum, V; Jonker, C

• Publication status: Published
• Peer review status: Peer reviewed
• Publisher: Association for Computing Machinery
• Host title: Proceedings of the 14th International Conference on Autonomous Agents and Multiagent Systems
• Journal: Proceedings of the 14th International Conference on Autonomous Agents and Multiagent Systems
• Pages: 473-481
• Publication date: 2015-12-31
• Acceptance date: 2015-05-03
• ISBN: 9781450334136
• Keywords: SBTMR