Random-fuzzy programming based bidding strategies for generation companies in electricity market environment

Generation companies are most concerned about that how to build optimal bidding strategy in order to maximize profits in a competitive electricity market. Building bidding strategy through estimating and simulating the uncertain information in market is one of the main methods to solve the problem. In wholesale competition electricity market which demand sides participate in demand sides and generation companies bid price at the same time. So the premise of building the optimal bidding strategy for generation companies is how to estimate and simulate the uncertain information such as the bidding strategy of competitors and demand sides exactly. At the time of estimating the uncertain information, generally based on historical trading data probability method is used for simulation. But it depends on statistical characteristics of data thoroughly. When historical trading data is insufficiency and dissatisfies statistical characteristics, probability method is very limited in practice. Therefore, someone applies the fuzzy set theory method to the field of bidding strategy and puts forward a method for building optimal bidding strategy based on possibility theory. But random uncertainty can not be described by this method. So for solving the optimal bidding strategy problem better, based on a new creditability theory which fits for programming problem included multiple uncertain information this article describes the uncertain information in market with random-fuzzy variables and establishes the random-fuzzy chance constrained mathematical programming model aimed at maximal optimistic profit under the condition of given confidence level. And a fast and efficient method for solving the model that random-fuzzy simulation combines with hybrid intelligent algorithm is put forward. The hybrid intelligent algorithm is composed of improved neural network algorithm and advanced genetic algorithm. Finally, the calculation case shows feasibility and effectiveness of the method.