Jakub Brzostowski

Predicting partner's behaviour in agent negotiation

We propose an improved approach for modeling behaviours of negotiation partners and predictive decision-making based on this modelling. Our prediction is based only on the history of the offers during the current negotiation. The mechanism estimates an influence of different factors contributing to partners behaviour during negotiation and uses this information to construct a prediction about agents future behaviour. The optimal sequence of offers is determined according to the prediction. The approach is tested in simple scenarios and the results comparing our approach to random strategy selection are illustrated.

e-Negotiation Systems and Software Agents: Methods, Models, and Applications

Negotiation is a decentralized decision-making process that seeks to find an agreement that will satisfy the requirements of two or more parties in the presence of limited common knowledge and conflicting preferences. Negotiation participants are agents who negotiate on their own behalf or represent the interests of their principals. When electronic negotiations enter the stage, these agents could be intelligent software entities that take part in the process of searching for an acceptable agreement. The degree of involvement of these “intelligent agents” in negotiations can range from supporting human negotiators (e. g. information search, offer evaluation) to fully automating the conduct of negotiations. Choosing the degree of involvement depends upon the characteristics of the problem in the negotiation. In this chapter, we review electronic negotiation systems and intelligent agents for negotiations. Different types of negotiation agents, their roles and requirements, and various methods for effective support or conduct of negotiations are discussed. Selected applications of intelligent negotiation agents are presented.

Adaptive Negotiation with On-Line Prediction of Opponent Behaviour in Agent-Based Negotiations

We propose an adaptive approach in agent-based negotiation involving on-line prediction of the opponent behaviour based on the parametric non-linear regression analysis. The predictive decision-making mechanism for the negotiation agent is based on the history of offers in the current negotiation encounter. In comparison to the related work the proposed approach allows the negotiation agents to predict more complex behaviour of the negotiation opponent in terms of mixture of its time-dependant and behaviour-dependant tactics. We perform experiments in order to validate the proposed approach. The results show that the predictive decision-making gives better results in terms of the utility gains for the adaptive negotiation agent as compared with a range of non-predictive negotiation strategies.

On possibilistic case-based reasoning for selecting partners in multi-agent negotiation

The paper proposes an approach for selecting partners in multi-agent negotiation with the use of possibilistic case-based reasoning It predicts the possibility of successful negotiation with other agents based on their past negotiation behaviour and the derived qualitative expected utility for the current situation The proposed approach allows the agents to select their most prospective negotiation partners based on a small sample of historical cases of previous negotiations even if they are different from the current situation Partner selection models for both independent and correlated negotiation agents are detailed and demonstrated with simple scenarios.

On possibilistic case-based reasoning for selecting partners for multi-attribute agent negotiation

We propose an enhanced mechanism for selecting partners for multi-attribute negotiation. The mechanism employs possibilistic case-based reasoning. The possibility of successful negotiation for each potential partner is predicted on the basis of its behaviour in previous multi-attribute negotiations. The qualitative expected utility for each potential partner is derived and the agents are ordered according to the values of these utilities. The order determines who is more and who is less desirable partner for negotiation. The proposed approach allows choosing the most prospective negotiation partners based on small sample of historical cases of previous interactions even if the previous situations are different from the current one. A simple example of calculations is presented to demonstrate the proposed approach.

Modelling partner’s behaviour in agent negotiation

The paper proposes new approach for modelling negotiation partners and predictive decision-making. It is based on the prediction of negotiation partners behaviour from its previous offers in the current encounter. The approach allows the negotiating agent to asses different factors influencing other agent’s behaviour during negotiation and make optimal decisions according to the prediction. It is tested in simple scenario and the results illustrating the comparison with random strategy selection are presented.

On Fuzzy Projection-Based Utility Decomposition in Compound Multi-agent Negotiations

In the process of compound multi-agent negotiation a number of agents concurrently negotiate with one or more counterparts in order to satisfy the individual preferences that lead to the collective maximization of the overall utility function imposed on the compound service. In order to perform this task the overall utility function has to be decomposed into individual single-service utility functions. This problem is not trivial, especially in compound multi-agent negotiations involving more complex aggregation patters of negotiated issues. In this paper we propose an approach for derivation of the individual utility functions based of the principles of fuzzy set projection. We also propose a way of modifying the initially generated utility functions in the case where the agreement was not reached with those functions, what allows for reaching an agreement in repeated negotiation.

Negotiation partners selection mechanism based on context-dependent similarity relations

This paper proposes a context-dependent case-based mechanism for selecting negotiation partners with the focus on the adaptation of similarity relations to a specific context. The similarity relations are important part of the reasoning mechanism and have to be defined in a way consistent with the specific type of interaction, the negotiation. We validate the proposed approach with the use of both the probability and possibility distributions by performing experimental evaluation.

Supporting Ill-Structured Negotiation Problems

The negotiation is a complex decision-making process in which two or more parties talk with one another in afford to resolve their opposing interests. It can be divided into consecutive stages, namely: pre-negotiation phase involving structuring the problem and the analysis of preferences, the intention phase involving the iterative exchange of offers and counter-offers, and the postoptimization phase aiming at the improvement of the agreement obtained in the intention phase. In this chapter, we focus on the analysis of negotiators\({^\prime }\) preferences in ill-structured negotiation problems. We employ the modified FTOPSIS approach and the AHP method for determining the negotiation offers\({^\prime }\) scoring system, which allows for the easy evaluation of both the incoming offers as well as the packages under preparation. The imprecision and vagueness of the packages and option\(\text {s}{^\prime }\) descriptions is modeled by the fuzzy triangular numbers. The Analytic Hierarchy Process is used to derive the negotiation issue weights instead of directly assigning such values to the issues (a classic approach). The FTOPSIS method is used to build the final scoring system allowing for the evaluation of any potential negotiation package. The whole process of negotiation supported by the approach we proposed is illustrated with an numerical example.

Experimental Evaluation of Possibilistic Mechanism for Negotiation Partners Selection

Negotiation is an interaction allowing agents to resolve conflicts and reach agreements over shared concerns [1]. The shared concerns of negotiation may be issues such as price or response time. Negotiation has been studied in different fields including management, social sciences, decision and game theory, artificial intelligence and intelligent agents [2]. Most of the literature devoted to negotiation considers the decision-making during the negotiation in terms of choosing appropriate negotiation strategies. However, there are some decisions that have to be made before the negotiation starts. One of such decisions is the selection of negotiation partners that typically is made by the agent’s user. The selection of negotiation partners is crucial because negotiation may be a very time-consuming activity and failed encounters can waste time and resources [3]. The selection problem may be solved by predicting the negotiation capability of each potential partner and choosing the required number of partners with the highest chance of success in a potential negotiation. The selection mechanism is very important because of the practicality and efficiency of multi-agent system interactions. Some research related to the problem of negotiation partners selection has been performed from the point of view of trust and/or service reputation (e.g. [4, 5]). In that work the credibility and commitment of the potential partner are taken into account in the agent selection process. However, other aspects like the preferences of a potential partner may have a strong influence on the outcome of potential negotiation. Another work by Banerjee and Sen [6] considers the problem of coalition formation. The agents decide which partnership to join for a fixed number of interactions based on the expected payoffs gained over a period of time. In that approach the consequences of action are modelled by a probability distribution and a classical notion of expected utility is used to asses the expected benefit of joining the coalition. However, that work does not consider negotiation but a different kind of cooperative encounter. Moreover, construction of the probability distribution in this ap-