Norms for DLP agents working in a warehouse scenario

Abstract

We have studied the development of multi-agents that must collaborate within their working environment according to a set of norms. The agents used are defined and implemented in dynamic logic programming (DLP), using the LUPS knowledge updating system developed in XSB. The aim of this study is to experimentally find out how norms imposed on agents can lead to an acceptable and robust collective behavior. Such a normative generated behavior is more natural in an open environment where new agents may enter the world with expectation of social acceptance. The experiments carried out have shown where we have flaws in the initial definition of the norms. We have been able to learn how to modify the norms in order to improve the social behavior of the group of agents in the warehouse scenario. © 2001 Springer-Verlag Berlin Heidelberg.