The Research On Bidding Strategies Of Generation Companies Based On American Put Option Contracts

PCR(principal component regress)that with the characters of reducing dimensions effectively and overcoming the intense relativity between independent variables, is widely used in different fields. Approximate programming is a good algorithm in solving non-linear optimizing question.This thesis evolves with principal component analysis and risk revenue ratio,considers two electricity markets,adopts PCR to seek option price and uses approximate programming to determine the optimal allocation of energy for generation companies between option market and spot market. First,the optimal stopping theory and the solution of Shell package are shown.A new method of simulating price for American put options,which is based on principal component analysis,is presented.This method is better than the least square method,because in the latter method an assumption that American put options can't be executed in advance.On the other hand,PCR can draw efficient information as much as possible,which can make the simulation converge at the exact result. Secondly,an incomplete information model which considers the unit cost and the maximum bidding price of power generation companies,the probabilities of winning below or on the margin has been built.Then a strategy based on expected MCP and Bayesian game theory for maximizing profit of generation companies has been presented. Finally,while considering price risk and optimizing profit,a dual model based on option market and spot market where MCP is influenced by option price is built and a method based on approximate programming of optimal total energy and allocation for generation companies is presented.