An agent-based electronic job marketplace: conceptual foundations and fuzzyMAN prototype

Abstract

In this paper, conceptual foundations and the architecture of an agent-based electronic job marketplace called FuzzyMAN (Fuzzy Multi-Agent Negotiations) are presented. Software agents on that marketplace negotiate about multiple issues according to a multilateral negotiation protocol that is described in detail in this paper. First, we define negotiation issues, a negotiation protocol, and negotiation strategies for a bilateral negotiation model. Based on the bilateral model, a protocol for multi-lateral negotiation with multiple negotiation issues is introduced. This protocol is based on a sequence of bilateral negotiations and on a pre-selection procedure that chooses suitable partners for negotiation first. Special attention is given to the feature that the agents' preferences are expressed in fuzzy terms. Although this paper deals with the specific situation of an electronic job marketplace, the proposed negotiation model can he used with appropriate modifications for other agent-based e-marketplaces where agents have contradictory aims and negotiation is required.