Autonomous and heterogeneous software agents must agree on how they will operate together, with a view to achieving individual and global goals. Explicit, machine-processable representations of the terms for joint work among agents --aptly named contracts --are thus required. We explore a rule-based formalisation for contracts: rules capture conditional norms describing situations arising during the enactment of a multi-agent system, and norms resulting from these situations. However, such rules may establish conflicting norms, that is, norms which simultaneously prohibit and oblige (or prohibit and permit) agents to perform particular actions. We present a mechanism to detect and resolve normative conflicts in a preemptive fashion: these mechanisms are used to analyse a contract and suggest “amendments” to the clauses of the contract. These amendments narrow down the scope of influence of norms and avoid normative conflicts. Agents propose rules and their amendments, leading to a contract in which no conflicts may arise.
|Title of host publication||Artificial Intelligence Research and Development, Proceedings of the 12th International Conference of the Catalan Association for Artificial Intelligence, CCIA 2009, October 21-23, 2009, Vilar Rural de Cardona (El Bages), Cardona, Spain|
|Editors||Sandra Sandri, Miquel Sànchez-Marrè, Ulises Cortés|
|Number of pages||10|
|Publication status||Published - 2009|
|Name||Frontiers in Artificial Intelligence and Applications|