Work allocations governed by social profiles for large scale heterogeneous infrastructure

Abstract

This paper outlines research in to resource allocation using software agents where allocations are influenced by social and economic backgrounds. Furthermore, it explores developing existing software agents to provision a multiple-agent system (MAS) where each agent is described and created based on a single text file and where the addition, reconfiguration and removal of agents is possible. Several heterogeneous computer resources exist, each having different backgrounds, origins and motivations. Several large scale computing challenges of a humanitarian, commercial or scientific nature also exist. The problems arising in the assemble of resources to address such challenges are distributed and decentralized in nature. Software agents have become a mainstream technology for addressing such problems by modeling organisation and the interactions within them. Agents intending to engage in such social behaviour need a mechanism to consider social, economic and cultural backgrounds as this is an initial step in formation of organisations. In this work, a textual description of an entity's social and economic outlook will be called a social profile. Search technology will be used within agents to differentiate such descriptions. This text file will also contain the agent configuration details, but will also be extensible to allow future developments.