Towards a new supporting platform for collaboration in industrial diagnosis within an agent-based WEB DSS

Abstract

The main objective of this study is to provide a professional social network which is integrated within an agent-based WEB DSS. This DSS network will facilitate interaction, discussion, and the sharing of information among production operators, especially to deal with nominal situations of resource faults. The coordination among the agents in our approach is made possible by a dynamic agent coordination protocol. The general architecture is based on agents named: production agent (PA), ontology agent (OA), evaluator agent (EA), and coordinator agent (COA). The coordination protocol that is applied comprised a set of behaviours used to maintain and control the messages exchanges among the agents. Learning features enable these agents to gain more time in calculation and executions. As a representation of the agent's common knowledge, domain ontology has been developed to represent major generic concepts in the industrial domain. Analytical hierarchy process (AHP) methodology is applied to evaluate sorting solutions in respect of human participants' preferences. In this study, we choose as a field of application: ALFATRON electronics industry.