Intelligent Workforce Allocation within an Agent-based Paradigm: Central and Distributed Decision Powers

B. Virginas, M. Ursu, G. Owusu, and C. Voudouris (UK)

Keywords

E-Commerce and Resource Management, Software Agents, Planning and Scheduling

Abstract

This paper presents an agent-based solution for a resource allocation problem within British Telecommunications (BT) plc. The solution is part of ARMS (Automated Resource Management System), an integrated system for customer service operations in BT. The generic paradigm subscribed by the solution is an agent-based model in which the decision power of each agent is made into an explicit feature. A variety of organisational structures can be obtained from the generic model, by assigning different decision powers to agents. The focus of our work is to analyse the behaviour of such organisational structures in specific problem-solving contexts. We have built two functional systems on the basis of two "extremes" of the model: a centralised system and a completely and uniformly distributed system. The centralised system is currently ready to go on trial. With the distributed system we are currently running experiments based on designed data sets. The results from these implementations lead to refinements of the generic model, which, in turn, constitutes the basis for further experimentation. Ultimately, they will converge into a unitary system. We believe that the integrated system will provide a better solution to workforce allocation within BT, on the one hand, and, furthermore, that it will constitute a toolkit reusable in other resource allocation contexts.

Important Links:



Go Back