An agent-based negotiation platform for collaborative decision-making in construction supply chain

Abstract

Negotiation is an effective and popular decision-making and coordination behavior in inter-organization systems, especially in construction supply chain (CSC) which is characterized by fragmentation, low efficiency and multiple partners. This chapter defines the concepts of CSC and CSC management, especially regarding CSC management as the coordination of inter-organization decision-making in CSC and the integration of key construction business processes and key members involved in CSC. Much research and practice indicates that there are still many problems in construction, most of which are supply chain problems. The research analyses the problems in CSC. In order to resolve these problems and improve the performance of construction, this research proposes a multi-agent based negotiation platform for improving the effectiveness and efficiency of collaborative decision-making in CSC by adopting agent technology and regarding CSC as a typical multi-agent system. The general structure of the agent-based negotiation platform (ANeP) is designed, which includes two kinds of agent group: specialty agents and service agents. Since different members in CSC have different preferences on the decision attributes (such as cost, time, quality, safety and environment), a multi-attribute negotiation model is established by designing a negotiation protocol and describing the negotiation process. Considering the negotiation failure or man-made termination of negotiation for increasing the efficiency of negotiation, this research presents a relative entropy method for improving agent-based multi-attribute negotiation efficiency (REAMNE) in ANeP. This method aggregates the preference information of negotiators in CSC twice. Firstly, compromise group preference order is ascertained by using preference information and compromise preference information. Secondly, group preferences are aggregated by using a relative entropy model, which is established based on entropy theory whilst, meanwhile, considering the multiple attributes in CSC negotiation. The prototype of ANeP is developed based on the ZEUS agent development toolkit. The trial run reveals the feasibility of implementing ANeP for improving collaborative working in CSC. © 2009 Springer-Verlag Berlin Heidelberg.