It is convenient for agents participating in computational institutions to have as much autonomy as possible. The model of electronic institution considered in this paper imposes restrictions on the communication between agents through a precise protocol. Otherwise the agents are reasonable autonomous. To foster some kind of acceptable social behaviour norms are used. When conflicting goals appear in the institution, their preference is expressed by norms that show the reward/punishment applied to agents obeying/disobeying a given norm. Several types of agents are defined: social, rebellious, selfish. Their influence on the performance of the institution is captured by utility measures expressing social contribution and individual satisfaction.
|Number of pages||19|
|Publication status||Published - 2004|