SG3: Norm-governed interaction in organisations

Topics

The application of automated information-processing techniques to an ever increasing number and variety of activities of organisations (both public and private) creates an urgent need for improved formal models of agent-interaction, and of the norms which are intended to govern that interaction. Whether the agents involved are humans, or computer systems, or the members of hybrid systems in which human and artificial agents both play their parts, the problems involved in formally characterising their organised interaction are considerable. Formal techniques that will be used in characterising this interaction will include action logic, deontic logic and non-monotonic logics.

Our focus will be on the following topics, each of which takes up some central aspects of organisational behaviour:

1. typically, some agents within an organisation are granted powers to authorise changes of various kinds. We shall investigate ways of representing such powers, and their associated rights, using tools of deontic logic and the logic of action. We also seek integration of the analysis of power within the theory of normative positions.
2. some business transactions involve interaction between distinct organisations, each of which may have its own procedures. We shall investigate the particular characterisation problems which arise when different norm-systems interact in such transactions.
3. Characterising features of modern integrated applications (EDI, just-in-time, ...) are that (1) we have a network of communicating agents, who (2) are subject to behavioural norms and who (3) act in real time. Such applications require a specification language with deontic and real-time features.

Goals

• The preparation of tutorials in which ModelAge members can be given a picture of the state of the art with regard to formal modelling in the above-mentioned problem areas.
• The identification of specific, concrete examples, from the domains of business and legal transactions, in which the characterisation/knowledge representation issues mentioned above are clearly exhibited. These may function as test cases for the adequacy of the proposed formal analyses.
• To relate work under SG3 to the study of revision (carried out under SG1), and to the logical modelling of action and interaction (carried out under SG2).

Schedule

SG3 will hold a yearly meeting. Work has already begun at the Workshop on Deontic and Non-MonoTonic Logics held at Rotterdam (Oct 4-5, 1993), where the revisable combination of norms (a topic shared with SG1) was considered.

The initial period, leading to the first main workshop, will focus on the preparation of tutorials to be presented at that workshop. Smaller, preparatory meetings between participants may be held before that time, as funds permit.

The second and third years should focus on further refinement of the emerging formal models, in the light of the test cases identified within 2, and in relation to the methods exhibited in SG1, SG2.