首页|The complexity of bribery and control in group identification
The complexity of bribery and control in group identification
扫码查看
点击上方二维码区域,可以放大扫码查看
原文链接
NETL
NSTL
Springer Nature
The goal of this paper is to analyze the complexity of constructive/destructive bribery and destructive control in the framework of group identification. Group identification applies to situations where a group of individuals determine who among them are socially qualified. We consider consent rules, the consensus-start-respecting rule, and the liberal-start-respecting rule. Each consent rule is characterized by two positive integers s and t, and the socially qualified individuals are determined as follows. If an individual qualifies herself, then she is socially qualified if and only if there are in total at least s individuals qualifying her. Otherwise, she is NOT socially qualified if and only if there are in total at least t individuals disqualifying her. The liberal (resp. consensus)-start-respecting rule determines the socially qualified individuals recursively. In the first step, all individuals qualifying themselves (resp. qualified by all individuals) are socially qualified. Then, the procedure recursively adds individuals who are not socially qualified but are qualified by at least one socially qualified individual into the set of socially qualified individuals until no one can be added this way.
Computational social choiceAlgorithmsGroup identificationBriberyControlComplexity
Gobor Erdelyi、Christian Reger、Yongjie Yang
展开 >
School of Mathematics and Statistics, University of Canterbury, Christchurch, New Zealand
School of Economic Disciplines, University of Siegen, Siegen, Germany
Economic Theory, Saarland University, Saarbruecken, Germany,School of Computer Science and Engineering, Central South University, Changsha, China