Towards a Quality Assessment Method for Learning Preference Profiles in Negotiation

Abstract

Abstract. In automated negotiation, information gained about an opponent's preference profile by means of learning techniques may significantly improve an agent's negotiation performance. It therefore is useful to gain a better understanding of the factors that influence the quality of learning. The benefits of learning in negotiation are typically assessed indirectly by means of comparing the utility levels of agreed outcomes and other more global negotiation parameters. An evaluation of learning based on such general criteria, however, does not provide any insight into the influence of various aspects of negotiation on the quality of the learned model itself. The quality may depend on such aspects as the domain of negotiation, the structure of the preference profiles, the negotiation strategies used by the parties, and others. To gain a better understanding of the performance of proposed learning techniques in the context of negotiation and to be able to assess the potential to improve the performance of such techniques a more systematic assessment method is needed. In this paper we propose such a systematic method to analyse the quality of the information gained about opponent preferences by learning. The method includes measures to assess the quality of a learned preference profile and proposes an experimental setup to analyse the influence of various negotiation aspects on the quality of learning. We apply the method to a Bayesian learning approach for learning an opponenent's preference profile and discuss our findings.

Towards a Quantitative Concession-Based Classification Method of Negotiation Strategies

“…The specifics of a negotiation domain can be of great influence on the negotiation outcome [8]; therefore, negotiation characteristics such as concession rate have to be assessed on negotiation domains of different size and competitiveness (or opposition [9]). With this in mind, we aimed for two domains (with two preference profiles each) with a good spread of negotiation characteristics.…”

A Multi-Agent Environment for Negotiation

“…It provides a variety of tools to analyze the performance of agents and may also be used to compute quality measures related to e.g. the quality of an opponent model [15]. The architecture that is introduced here identifies the main integration points where adapters are needed to connect a negotiating agent to this architecture.…”

“…The higher the value of the ranking distance the stronger opposition between the preference profiles. Another measure for the opposition of preferences proposed in [15] uses Pearson's correlation coefficient for that purpose. This coefficient represents the degree of linear relationship between two variables.…”

Predicting the Performance of Opponent Models in Automated Negotiation

“…Pearson correlation of issue weights [6] Pearson correlation coefficient between real and estimated issue weights. Ranking distance of issue weights [6] Ranking distance between real and estimated issue weights.…”