Generating Pareto Solutions in a Two-Party Setting: Constraint Proposal Methods

Abstract

This paper presents a class of methods, called constraint proposal methods, for generating Pareto-optimal solutions in two-party negotiations. In these methods joint tangents of the decision makers' value functions are searched by adjusting an artificial plane constraint. The problem of generating Pareto-optimal solutions decomposes into ordinary multiple criteria decision-making problems for the individual decision makers and into a coordination problem for an assisting mediator. Depending on the numerical iteration scheme used to solve the coordination problem, different constraint proposal methods are obtained. We analyze and illustrate the behaviour of some iteration schemes by numerical examples using both precise and imprecise answers from decision makers. An example of a method belonging to the class under study is the RAMONA method, that has been previously described from a practical point of view. We present the underlying theory for it by describing it as a constraint proposal method, and include some applications.negotiation analysis, Pareto optimality, joint problem solving, multiple criteria

A decentralized model for automated multi-attribute negotiations with incomplete information and general utility functions

“…There exist three pieces of work [6,12,29] that introduce a non-biased mediator into the system to help agents negotiate in complex situations. In [6], Ehtamo et al present a constraint proposal method to generate Pareto-frontier of a multi-attribute negotiation with adopting a non-biased mediator.…”

“…In [6], Ehtamo et al present a constraint proposal method to generate Pareto-frontier of a multi-attribute negotiation with adopting a non-biased mediator. The mediator generates a constraint in each step and asks agents to find their optimal solution under this constraint.…”

“…[2,3,19]), agents are cooperative, or a non-biased mediator is available (e.g. [6,12,29]). Thus, the absent in the prior work is a protocol that can not only assist agents efficiently to make offers in an n-dimensional space with incomplete information but also give self-interested agents sufficient decision flexibility, for instance, the right to select offers to make and the right to accept a given offer.…”

Agent Reasoning in Negotiation

“…(Ehtamo et al 1999) present a constraint proposal method to generate Paretofrontier of a multi-attribute negotiation. The mediator generates a constraint in each step and asks the agents to find their optimal solution under this constraint.…”

Comparison of decision rules to mediate negotiations with incomplete information

“…According to the "Dual Concern" model (Pruitt, 1983;Thomas, 1992), solutions that improve the position of the parties relatively the current situation only can be achieved if negotiators have enough knowledge about their preferences and about the preferences of the other parties. Ehtamo et al (1999) present a class of methods called "Constraint Proposal Methods", which are interactive methods to find Pareto optimal solutions through common tangent hyperplanes, for negotiations between two parties about two or more continuous issues, with the help of a mediator. Heikanen (1999) proposes an interactive method to determine Pareto optimal solutions in negotiations with multiple parties about continuous issues, using a mediator.…”