Multi-Agent Systems are computational systems in which a collection of autonomous agents interact to achieve a certain task, for example to fulfil an obligation directed to the whole group, i.e., a collective obligation. Since, such a collective obligation is beyond the capacity of an individual agent, the agents have to communicate, cooperate, coordinate and negotiate with each other, to achieve the collective task: the fulfilment of the obligation. In this paper we discuss and formalise collective aspects of obligations and commitments. Collective obligations are analysed and formalised in a deontic logic framework. The notions of individual and collective commitment are defined to specify which individual has the responsibility to fulfill an ‘internal’ obligation as part of the collective obligation. In distributed artificial intelligence (DAI) theories of organisations, it is emphasized that ‘commitment’ is a crucial notion to analyse a collective activity or the structure of an organisation. In this paper we give a first attempt to formalise the notion of commitment to determine which plan has to be followed to achieve a joint goal, i.e. the fulfillment of a collective obligation by using several concepts as commitment, delegation and authority-relation.
|Title of host publication||Database and Expert Systems Applications (Proceedings 11th International Conference, DEXA 2000, London, UK, September 4-8, 2000)|
|Editors||M. Ibrahim, J. Küng, N. Revell|
|Place of Publication||Berlin|
|Publication status||Published - 2000|
|Name||Lecture Notes in Computer Science|