Supplier multi-trading strategy: A stochastic programming approach

A power supplier in deregulated environment needs to allocate its generation capacities to participate in contract and spot markets. The well-known mean-variance method is inappropriate to deal with assets whose price distribution is non-normal. In order to model the electricity assets with different distributions into portfolio optimization, this paper proposes a stochastic programming approach based on genetic algorithm and Monte-Carlo simulation. In the real market data based numerical study, the performances of the proposed method and the standard mean-variance method are compared. It was found that the proposed method can obtain significantly better portfolios in the situation that non-normally distributed assets exist for trading. The modeling capacity, flexibility and robustness will make the proposed method potentially useful in application.