Paper Info
Title
A Region-Based Multi-Issue Negotiation Protocol for Nonmonotonic Utility Spaces
Abstract
Nonmonotonic utility spaces are found in multi-issue negotiations where the preferences on the issues yield multiple local optima. These negotiations are specially challenging because of the inherent complexity of the search space and the difficulty of learning the opponent's preferences. Most current solutions successfully address moderately complex preference scenarios, while solutions intended to operate in highly complex spaces are constrained by very specific preference structures. To overcome these problems, we propose the Region-Based Multi-issue Negotiation Protocol (RBNP) for bilateral automated negotiation. RBNP is built upon a nonmediated recursive bargaining mechanism which efficiently modulates a region-based joint exploration of the solution space. We empirically show that RBNP produces outcomes close to the Pareto frontier in reasonable negotiation times, and show that it provides a significantly better performance when compared to a generic Similarity-Based Multi-issue Negotiation Protocol (SBNP), which has been successfully used in many negotiation models. We have paid attention to the strategic issues, proposing and evaluating several concession mechanisms, and analyzing the equilibrium conditions. Results suggest that RBNP may be used as a basis to develop negotiation mechanisms in nonmonotonic utility spaces.
Year
DOI
Venue
2011
10.1111/j.1467-8640.2011.00377.x
COMPUTATIONAL INTELLIGENCE
Keywords
Field
DocType
multi-agent systems,automated negotiation,multi-issue,nonmonotonic utility spaces
Mathematical optimization,Computer science,Local optimum,Knowledge management,Equilibrium conditions,Multi-agent system,Artificial intelligence,Recursion,Pareto principle,Machine learning,Negotiation
Journal
Volume
Issue
ISSN
27.0
2.0
0824-7935
Citations 
PageRank 
References 
11
0.64
26
Authors
4
Name
Order
Citations
PageRank
Miguel A. Lopez-Carmona117519.82
Ivan Marsá-maestre29912.15
Enrique de la Hoz311815.35
Juan R. Velasco431936.36