Seleção de portfolios de projetos de produção de petroleo por intermedio de tecnicas de otimização e curvas de isopreferencias / Portfolio selection of petroleum production projects using optimization techniques and indifference curves

AUTOR(ES)
DATA DE PUBLICAÇÃO

2005

RESUMO

The companies in the E&P business dedicate significant time and resources to decide the allocation of capital investment as a result of the large number of potential projects in their portfolios. The approach known as portfolio optimization aims to choose a subset ofprojects. The methodology used in order to achieve the portfolio optimization is based on maximizing return (NPV, IRR, etc.) and minimizing risk (NPV standard deviation, variance, etc.), according to technical and operational corporate goals of a determined company. The quantitative tools (optimizers) used in this dissertation to achieve portfolio optimization were: linear programming, random programming and genetic algorithms. However, those approaches do not take into account explicitly the risk attitudes of the corporate decision makers. In order to overcome these constraints, the attitude front to the risk of two decision makers of a petroleum E&P company was modeled using preference theory. The main objective of this dissertation was to develop an integrated methodology for portfolio optimization and selection of petroleum production projects using optimization techniques, portfolio modern theory and preference theory. This extended model was applied to a set of 25 heavy oil offshore production projects located in Brazilian area. Seven corporate strategies were used to reproduce a number of representative portfolios of the most common conditions in this environment. These portfolios allow different risk and return values. The decision maker s attitude towards risk was inc1uded into the portfolio selection model by using functions which reflects a choice under an uncertainty environment. The results achieved indicate that the genetic algorithms optimizer generates portfolios which are concerned to the proposed objectives and constraints, while the random generator optimizer generates portfolios which don t respect corporate constraints and goals. The application of the preference models has presented more consistent results in the portfolio selection process, because it allows incorporating the attitudes of the individuaIs front to the amount of capital exposed to the risk

ASSUNTO(S)

uncertainty risco programação linear genetic algorithms linear programming monte carlo method risk algoritmos geneticos metodo de monte carlo incerteza

ACESSO AO ARTIGO

Documentos Relacionados