Study Of Stochastic Distribution Model Of Gencos' Bidding Behaviors In Electricity Market

Since 80s of the 20 century, the electrical power industry reform bandwagon had swept across the world and a series of marketing reform followed naturally. As the introduction of competition in electricity market, Generation companies (Gencos) have opportunities to make various bidding strategies based on their own power generation cost. In order to maximize their profit, they may withhold power and collude with other Gencos in submitting their bid price quotations, which result in the shortage of market capacity and a rise in the market clearing price ,even soaring price, therefore, it is quite essential for supervisions to conduct and further study on Gencos'bidding behaviors and master variation regularity of their bidding behaviors and reveal abnormal bidding behaviors caused by human element.Gencos'bidding behaviors are affected by human element and outside factors. Although the leading cause of Gencos'abnormal bidding behaviors and soaring price is human element, outsider factors couldn't be neglected or wrong results will be obtained. The previous methods usually neglect the impact factors of Gencos'bidding behaviors, which make supervision and regulation department could not acquire adequate and reliable information to analyze overall Gencos'bidding behaviors caused by human element. To make study more practical and valuable, the study about Gencos'bidding behaviors should be conduct on the premise of that the identical influence of outsider factors on Gencos'bidding behaviors to eliminate the interference of outsider factors, moreover, these methods are restricted within in a narrow sphere with essentially a focus on some special indexes, while there are few profound studies in stochastic distribution of Gencos'bidding behaviors at present. To solve above problems, this paper clusters outside factors data in observing periods to obtain the classes of influence of outsider factors on Gencos'bidding behaviors firstly, from the view of power supervision, then proposes stochastic distribution model Gencos'overall bidding behaviors under each class. Considering the fuzziness and uncertainty of outsider factors, the paper applies a new mathematic tool called fuzzy soft sets which could handle fuzzy and uncertain problem well to cluster outside factors data, then estimates one dimension stochastic distribution of Gencos'bidding behaviors based on support vector machines which could solve small-sample, non-linear and high dimension problems in machine learning, and through analyzing bid data, this paper also expands one dimension stochastic distribution model to two dimensions. The supervisions could analyze bidding behaviors and conduct effective monitoring with a clear aim according to the stochastic distribution. Finally, a numerical example is used to illustrate the feasibility and applicability of the proposed method. The calculation results show the proposed model can truly reflect the characteristics of distribution under different influence of outside factors and provide supervision of whole market a new way for security monitoring and risk prevention.