The New Models Of Market Forecasting And Bidding Decision Optimizing For The Power Generation Company

The paper researches the problem that how power generation companies effective indentify and manage the potential market risks based on electricity market forecasts, and avoid the loss, thereby enhance the benefit ability from competitiveness. The paper masters the core of electricity market ---MCP, selects advanced artificial intelligence combination forecasting methods to model and predict, thus forms a precise prediction analysis system for generation companies in short-term market environment. Then, the paper builds intelligent optimization model from the aspects such as benefit, MCP, generation and so on, based on group decision-making algorithm optimization. On this basis, draw on the theory of financial market risk management, the paper quantitatively describes the risks in the electricity market. Combined with competitive risks of power generation company, the paper quantities the assessments. On the mentioned above, the paper proposes the bidding decision-making optimization model for generation companies with a risk-control, and design programming optimization algorithm.The main foundation the paper relied on is optimized combination forecast theory, intelligent optimization computation theory, the time series forecast, and financial risk measurement and management theory. This paper will unify the newest achievement of these fundamental researches applied to the content which this paper study, and aim at the special questions to improve the relative theory models and develop the connotation and the extension of these theory, and then form a complete set of the electricity generation companies bidding decision-making optimization theory based on MCP and short-term market risks, and provide power generation companies with theory instruction and special analysis tools for benefit maximization and risk avoidance.The innovation of the paper is reflected in the following aspects:1. A new load forecasting model with single intelligent is built. Wavelet neural networks improved by genetic algorithm, fuzzy chaotic neural network model are applied to predict short-term power load. Through introducing new load factors and data mining, the paper constructs the new factors indicator set. By introducing the thought of optimal combination forecasting, based on the idea that the improvement of the ant-colony-optimization, the paper constructs load combination forecasting model with intelligent dynamic empower, and analyzes the sensitivity of affecting factors; 2. Based on achievement of comprehensive time series forecasting and neural network prediction, the paper proposed price forecasting model of mixed time series based on support vector machines. Based on statistical methods and intelligent pattern recognition method to study price nails formation mechanism, forecasting methods, thus the right price mutation predicted the new data mining method;3. Bidding simulation and decision-making model of power generation companies based on price and optimization algorithm design. Through assessing power generation company property in short-term, taking into account the risks, the bidding of the generating units optimization model and algorithm based on VaR method is proposed.