A Study On Bidding Assistant Decision Algorithm Used In Electricity Multilateral Trading Market’Offer

Electricity price is taking centre stage in the resources flow and the distribution of interests under the environment of power market. All parties involved in the market transact business with each other on the basis of electricity price. With the development of Inner Mongolia electric power multilateral trading market, power generalization enterprises are in poor state of studying the algorithm of bidding decision prediction in order to gain greater economic benefits.This paper aims to study the algorithm of bidding decision prediction which can be applied in the Inner Mongolia Electric Power multilateral trading market on the basis of its operational rules. In the first part of this paper, the advantage and disadvantage of Regression forecast method、Particle swarm optimization method、Gray algorithm、genetic algorithm、Neural network algorithm and a combination of all these methods have been mainly analyzed and compared. In the following part, the assistant decision algorithm of three kinds of quoted market price which are the algorithm of multiple linear regression, Generalized neural network and support vector machine have been used and the historical transaction data of the Inner Mongolia Electric Power multilateral trading market has also been utilized to emulate these algorithms. Multiple linear regression decision algorithm is suitable for situation when the decision price difference changes linearly, but not suitable for situation when it changes in needle pattern. In contrast, the algorithm of generalized neural network aided decision and support vector machine is fit for situation when decision price difference changes in needle pattern but the algorithm doesn’t fit for when decision price difference changes linearly. Besides, generalized neural network aided decision algorithm produces better result when the difference of electrovalency changes significantly.In opposite, the support vector machine algorithm shows better result when the difference of electrovalency changes small. Next, this paper put forwards some combination strategies based on the Time-varying Weighted Combination algorithm when it deal with the problem of the accuracy of electricity prediction which becomes low on some special circumstances. The result of emulation shows that combined assistant decision algorithm is able to reduce the impact of changes of various factors in the market operation, therefore it will increase the accuracies of quoted prices of electricity generation enterprises during their participation. In conclusion, this paper deals with problems of assistant decision algorithm which arise during the participation of electricity generation enterprises in the Inner Mongolia Electric Power multilateral trading market based on the its characteristics. Four kinds of bidding assistant decision algorithm have been put forward which will bring scientific references for the participation of electricity generation enterprises and optimizing their operations in order to achieve the goal of the participation of electricity generation enterprises about quoted market price and aid decision. The goal consists of two parts:First it will enhance the accuracy of the electricity market price forecast. Second, it will increase the profits of electricity generation enterprises. Based on the discussion above, bidding assistant decision algorithm of electricity generation enterprises studied and propound in this paper has great significance in both theory and practice.