Manipulação ótima de preferências no modelo de grafo para resolução de conflitos

Abstract

Conflicts are inherent in human relations and occur at various levels from the personal scale to conflicts involving large blocks of countries. The costs arising from conflicts are of various types: economic, social and environmental. Thus, effective ways of intervening in conflicts to achieve stability in certain desirable scenarios are of great importance. The Graph Model for Conflict Resolution (GMCR) is a model that has long been used to model and analyze conflicts because it is flexible and easy to calibrate. The purpose of this Master’s thesis is to present ideas on how to work with the inverse GMCR to optimize costs in changing the preferences of each DM to achieve equilibrium states within the conflict. We propose some methods to aggregate costs of changing DM’s preferences. The purpose is to determine the lower aggregate cost preference changes that make a given desired state an equilibrium according to a given stability notion. Besides formally describing the problem, we study some properties of the minimum costs for different stability notions, determine the computational complexity of this problem, and propose two algorithms for solving this problem in bilateral conflicts, one based on an exhaustive search, which turned out to be inneficient, and another method that is based on an integer linear programming problem. We apply this method to two known conflicts in the GMCR literature: the conflict of the Cuban missile crisis and the conflict of values.