Evaluation of Payoff Matrices for Non-Cooperative Games via Processing Binary Expert Estimations

• Main Author: Romanuke Vadim
• Format: Article
• Language: English
• Published: Sciendo 2016-12-01
• Series: Information Technology and Management Science
• Subjects: Estimation procedure; expert’s binary judgment; non-cooperative game; payoff averages; payoff matrice evaluation
• Online Access: http://www.degruyter.com/view/j/itms.2016.19.issue-1/itms-2016-0004/itms-2016-0004.xml?format=INT

A problem of evaluating the non-cooperative game model is considered in the paper. The evaluation is understood in the sense of obtaining the game payoff matrices whose entries are single-point values. Experts participating in the estimation procedure make their judgments on all the game situations for every player. A form of expert estimations is suggested. The form is of binary type, wherein the expert’s judgment is either 1 or 0. This type is the easiest to be implemented in social networks. For most social networks, 1 can be a “like” (the currently evaluated situation is advantageous), and 0 is a “dislike” (disadvantageous). A method of processing expert estimations is substantiated. Two requirements are provided for obtaining disambiguous payoff averages along with the clustered payoff matrices.