Forecast models and the nature of the decision process: The case of a Brazilian electricity distribution company

Abstract

Needs for adequate long-term planning in the execution of energy generation to meet growing demand leads to intense regulation on energy distributors. A research was realized by a Brazilian distribution company to identify the consumption prediction methods, the nature of the decision process and their specific interaction scheme. Were initially identified the models already used by other distribution companies. After, were studied the accuracy of the models and the models were adjusted based on the industrial consumption, with the use of Box-Jenkins methodology. Finally, were established criteria which have an impact on the decision process, using Herbert Simon's theory; demonstrating the limited value of any existing tool, and the importance of using more than one, integrating quantitative to qualitative analysis.