On The Extensive Form Of N-Person Cooperative Games

Abstract

This paper is concerned with improving our conceptual understanding of the extensive form of N-person cooperative games. The extensive form of N-person cooperative game is such that the game is played repeatedly for very much number of times, such that in the long run, the chances of being favoured and not being favoured are equally likely. Using the same examples as used by [3], it was observed that the Nucleolus method turned out to be better with standard error of 0.1334 and 0.2887 and coefficient of variation of 7.7% and 9.38% than the Shapley value method with standard error of 0.1498 and 0.3442 and coefficient of variation of 8.65% and 11.19% respectively. We also observe that in all, the standard error and coefficients of variation using both methods are lower in the extensive form of Nperson cooperative games than in the normal form of N-person cooperative games.

Keywords: Extensive form game, Normal form game, characteristic function, Coalition, Imputation, Player, Payoff, Strategy and Core