Abstract：Customarilymodern procurement adopts a multi-purpose procurement way, so the bidevaluation and contractor selection will be a great problem. The purpose ofthis paper is to provide a new method to resolve the problem by usingregression analysis based on facts-set. This method can evaluation all bidswith some cheap computer software. By proving the scientificity of this newmethod, the author thinks that it can avoid the defects of the past evaluationmethods efficiently. In virtue of the regression analysis to carry on the bidevaluation, this paper puts forward three kinds of different concrete ways:comparison of the residuals, using equivalent-benefit curve, setting upindexes-related evaluation system. The method can be used separately to tradeoff two or several factors, and can handle indexes whatever linear correlationor nonlinear correlation. This method also can be combined with other method inthe procession of bid evaluation.

Keywords：Regression analysis; procurement; Bid evaluation;Decision support; facts-set

  I. Backgrounds

Purchasingby invitation to bid of the present commercial and public ministers alwaysbases on multi-aims procurement, which don’t take price to be the only considereditem, but under the condition of fulfill the budget constraint and other basicrequirement to choose from a propriety of lower price, good quality and betterservice so etc evaluate items to arrange and form the best suppliers. In otherwords, that means purchasing the most worth services at an acceptable range ofprice, which not equals to the lowest price. 

Whichpair of the resultant could be the “best”? How to measure the competitors’advantage equations act on different aspects, with an order to get anintegrated result? Normally saying, every suppliers of bidding competitorscannot gain the championships in every item, such as better quality and morefunctions means for higher cost and price meanwhile; priority in price perhapsstands for less service. Especially in condition of modern times, all sorts ofenterprises will emphasis the individualized and characteristics developmentand extrude several parts’ advance. Therefore scientifically solve thetrade-off between each counting items and figures, which is a necessaryquestion for us to cross over in the way of gaining procurement aim and achieveefficiency optimization.