Research On Bid Evaluation Method Of Power Procurement Based On Support Vector Regression

In China, the Electric Power Industry is a foundamental industry which has something to do with the national economy and the people's livelihood, it takes on heavy responsibilities in providing the engine for development of national economy, it has ability to protect people's livelihood and maintain the stability of society. Since the reform and opening up, rapid development of China's economy had improved people's material and cultural living standards, also stimulated the demand for electricity consumption in China, promoted large-scale construction and development of China's power industry. Bidding is the primary means to get equipment and materials for power grid companies, and bidding evaluation is a core part of the bidding process, its rationality and fairness is directly related to the quality of equipment and operating efficiency of the grid company. Traditional evaluation methods more or less exist limitations. It is an important issue facing the evaluation organization to develop and apply the new evaluation method.The author has studied in the selection of evaluation methods in power grid's bidding, put over an evaluation model based on the theory of support vector machine.Firstly, The author has introduced bidding's theoretical background and the present situation of bidding in power industry, analyzed advantages and disadvantages of some traditional evaluation methods;Secondly, The author has elaborated the theory and principle of support vector machine, introduced solving procedure of support vector regression, described the basic idea of two optimization in parameters; then analyzed the applicability of support vector machine, put forward the idea of applying support vector regression in the evaluation process. Considering to the Characteristics in power biding, the author has established a rational and practical evaluation system as an example of supplies, analyzed meaning and quantization of each index in detail. Then established the evaluation model based on support vector machine.Thirdly,Took example for an evaluation in bidding of a power gid company, The author established the model, Using the grid search algorithm and particle swarm optimization to optimize the parameters, then get the optimal model for the case data, and tested and analyzed the learning performance of model under the two search parameters.Finally, the author simulated an evaluation in power grid's equipment procurement. And described the significance of the model's application in evaluation practice.Further, The author pointed out the inadequacies of the paper, and from another point of view, the author preliminarily studied in how to improve the model's learning and promoting ability, provided a new way of thinking for further study in the future.