Módulo de otimização para a determinação de compra e venda de energia elétrica em leilões

The commercial model adopted in the Brazilian Electric Sector since 2004 introduced a new process of purchase and sale of electricity starting from auctions. With the inclusion of customers in the free market, there was a need for studies and development of tools for the use of electricity contracting by the industry since a correct choice of energy volume and price can guarantee the company's profit or the viability of its service. In this sense, the present work presents a module based on open source programming with Java language for computational optimization of the data obtained from the generation of hydroelectric plants and their thermal complementation. Based on artificial intelligence technique, more specifically in the field of genetic algorithms, the computational tool allows to simulate different scenarios and future possibilities so that through an initial population it is possible to obtain several generations with the best characteristics of each individual and the amount of energy to be distributed between auctions. In this sense, it is possible to mitigate excess purchases and improve distribution to avoid buying below the necessary demand negotiated by the distributing company. For this, the return of a vector with the best selection of energy distribution in future auctions is then delivered to the end user in order to collaborate with their decision making. The results obtained with the reduction of over contracting and subcontracting of the purchase of electric energy demonstrated the feasibility of the program, allowing the use of this tool to assist agents in decision making that leads to the minimization of business costs involving energy. Finally, it can be concluded that the risks of future investments can be minimized.