Tomasz Szapiro

Genetic algorithms approach to a negotiation support system

It is argued that negotiation rules can be learned and invented by means of genetic algorithms. The work presented introduces a method, a system design, and a prototype implementation that uses genetic-based machine learning to acquire negotiation rules. The learned rules support a party involved in a two-party bargaining problem with multiple issues. It is assumed that both parties work towards a compromise deal. The method provides a framework in which genetic-based learning is applied repetitively on a changing problem representation. System design proposes a problem representation that is adequate to express bargaining processes and that is at the same time conducive to genetic-based learning. The authors report results of experiments with the prototype implementation. These results indicate that genetically learned rules, when used in real negotiations, yield results that are better than results obtained by humans in the same negotiation. The experiments indicate considerable robustness of genetically learned rules with respect to varying parameters defining the genetic operations on which the system relies in modeling negotiations. In terms of user support, experimental results show that in the bargaining process, a good rule is one that advises conceding in small steps and bringing new issues into the negotiation process. >

Generalized approach to modeling negotiations

Abstract Based on a concept of pressure which is more general and flexible than the commonly accepted utility approach, this paper develops a general approach for structuring modelling negotiations. The proposed model makes it possible to combine satisficing and optimizing approaches and to discover the best satisficing compromise which may be a dominated or even a non-vertex solution. The set of notions and definitions enables description of different types of negotiations and from different points of view. Two models of simple types of negotiations which are presented in the paper give a basis for further research on more complex types of negotiations.

On testing performance of a Negotiation procedure in distributed environment

Factors, which increase negotiation effectiveness challenge researchers to investigate methods supporting negotiation and to verify performance of negotiation support procedures. Increase of complexity of economic organizations as well as global range of their operation result in growing interest in negotiation involving remote parties coping with ambiguities of cross-cultural interactions. Trials to create distributed negotiation environment (DNE) answer the call for integrated set of tools, methods and procedures organized to support remote parties in negotiation. The DNE requires consideration of three functions of Distributed Negotiation Support System (d-NSS): communication, advice and mediation. Prototype d-NSS, which were presented in the literature, usually do not perform all three functions simultaneously. However, fully functional d-NSS should allow flexible switching between the three functions as well as to profit from the opportunity of their joint use. Such approach enables to reduce cost of tedious, expensive, time-consuming negotiation process. The paper reports on outcomes of a series of experiments dealing with DNE. The communication function was investigated in a distributed environment, initially asynchronous then synchronous. Advice and mediation functions were analyzed in experiments with well-structured data set. The results encourage efforts to include the issue of cautious use of a d-NSS in relevant cases as a topic in the pre-negotiation agenda.

A Review of Attribute Disclosure Control

Attribute disclosure occurs when the adversary can infer some sensitive information about an individual without identifying individual’s record in the published data set. To address this issue several privacy models were proposed with the goal of increasing the uncertainty of the adversary in deriving sensitive information from published data. In this chapter, firstly we review the underlying scenario used in statistical disclosure control (SDC) and Privacy-Preserving Data Mining (PPDM). In this chapter, we describe the attribute disclosure underlying scenario, the different forms of background knowledge of the adversary the adversary may have and their potential privacy attacks. then, we review the approaches introduced in the literature to tackle attribute disclosure attacks.

Outlooks and Insights on Group Decision and Negotiation

This book constitutes the proceedings of the 15th International Conference on Group Decision and Negotiation, GDN 2015, held in Warsaw, Poland, in June 2015. The GDN meetings aim to bring together researchers and practitioners from a wide spectrum of fields, including economics, management, computer science, engineering, and decision science.From a total of 119 submissions, 32 papers were accepted for publication in this volume. The papers are organized into topical sections on group problem structuring and negotiation, negotiation and group processes, preference analysis and decision support, formal models, voting and collective decision making, conflict resolution in energy and environmental management, negotiation support systems and studies, online collaboration and competition, and market mechanisms and their users.

Data Privacy: From Technology to Economics

In this paper, we attempt to relate two different approaches to data privacy – the technological approach, embodied in the current Privacy-preserving Data Mining work, and the economic regulations approach. Our main thesis is that none of these two approaches alone will be able to address the important, growing data privacy issues. We advocate a hybrid system, combining both approaches in a complementary manner. We present a view of privacy in the context of an accepted taxonomy of economic goods, stating the question: if privacy is exchanged and traded, then what kind of good is it? We also show that, viewed in the light of an established economic growth theory, involving privacy in the growth process leads to a steady state growth.

Penalty Rules in Multicriteria Genetic Search

In the paper, by means of numerical experiments conducted on artificially constructed problem instances, we test penalty rules for constrained genetic optimization of the Capacitated Heterogeneous Vehicle Routing Problem with Time-Windows in a bi-objective framework. Optimized criteria are cost minimization and capacity utilization maximization. Two approaches are employed – scalarization of objectives and dominance-based evaluation of solutions. We show that it is possible to handle infeasibility in such a way, that this risk of divergence to regions of infeasibility is acceptable. The most secure penalty rule among the tested ones turns out to be the rule which explicitly controls the proportion of infeasible solutions in the population. This rule, along with the rule which accounts only the notion of solutions distance from the feasible set, outperforms rules based on time-penalties and best to best-feasible solution comparison over considered case studies.

On Shaping Multi Criteria Marketing Strategy of a Pension Fund

In the paper the case study concerning multi criteria optimization of promotion campaign in a pension fund is considered. The significance of the problem is implied by the decisive role of marketing activities of firm which attempts to effectively compete in market environment. The impressive costs of TV campaign forces to focus on optimization of promotion. The specificity of pension fund is reflected in natural definition of the target group, in the structure of options and criteria and in evaluation procedures. Without loss of generality these procedures are simplified (the number of variables and criteria was left enough numerous to show multi criteria nature of the problem but restricted in order to clarify the presentation). The set of assumptions on a procedure supporting decision-makers was identified based on experience. It appears that the problem has a natural mathematical description which allows to formulate strictly the problem and to determine the set of solutions. The review of the set of solutions requires additional information. A procedure was presented which allows to collect the information needed. In order to collect this information the user is not forced to use any technical terminology but natural language. Collected data can be used to control the process of learning and elicitation of preferences. Randomly generated data illustrate this methodology which can be used for the real data as well.

Computer Based Analysis of an Economy in Transition to Steady State Equilibrium

We present a simple disequilibrium model of an economy in transition, which enables analysis of dynamic interrelations between interest rates, inflation, money and output. We also study effects of a change in nominal and real money on output.

A Safe Choice Procedure in Interactive Programming

In this paper we consider a multiple criteria decision problem framed as a vector optimization program (vop). Within this framework we propose an approach to problem resolution which allows for the selection of a compromise decision. This decision is sought by a decision maker (DM) through an iterative process, with subsequent proposals of the decision in question generated at different stages of this process. At each stage, a DM is asked to compare the decision being considered with the worst compromise, represented by the least acceptable levels of performance, and to express his opinion about the compared outcomes. This information is later used to generate a new decision, and the process is repeated. Following changes in a DM’S preferences the worst compromise is subsequently displaced.