An Agent-Based Model Applied to Brazilian Wind Energy Auctions

Marcos R. MacHado, Murilo Kenichi Fujii, Celma De Oliveira Ribeiro, Erik Eduardo Rego

Research output: Contribution to journalArticleAcademicpeer-review

3 Citations (Scopus)


This article, for the first time, adopts the agent-based model simulation technique to analyze the pricing process of energy in the Brazilian electricity market (auctions). Within this model, it is possible to analyze how the energy price is affected when a government intervention is observed through the increase in number of public companies participating in the auctions. In this paper, auctions of new and reserve energy of wind power are simulated. Through this model it is possible to compare the choice of bids from participating sellers in the auctions, categorized in two different groups: public and private companies. The agents (sellers) participate in the auctions by learning from the historical and simulated auctions that is regulated by the Brazilian government. Learning is performed through the usage of a variation of the Q-learning algorithm, which provides the sellers the optimal price-bid considering the conditions presented, which means that this price-bid will provide them the maximum reward possible. The results clearly show the average price difference between both generator profiles. In addition, it is possible to state that the price of energy changes due to the relative participation of public or private sellers in the auctions.

Original languageEnglish
Article number8891956
Pages (from-to)865-874
Number of pages10
JournalIEEE Latin America Transactions
Issue number5
Publication statusPublished - May 2019
Externally publishedYes


  • Agent-based
  • Energy pricing
  • Energy regulation
  • Wind energy auction
  • n/a OA procedure


Dive into the research topics of 'An Agent-Based Model Applied to Brazilian Wind Energy Auctions'. Together they form a unique fingerprint.

Cite this