Study On Mechanism And Models Of Multi-attribution Reverse Auction

E-procurement is an effective method for companies to save cost and develop efficiency. Recently, multi-attribute reverse auction (MRAT) has been adopted in E-procurement by lots of companies, the utilities of which using MRAT have greatly increased. In this paper, based on the deficiency of current research, applying mechanism design theory such as Revelation Principle (RP), a MRAT Mechanism is designed with Incentive Compatibility Constraint (IC), Individual Rationality Constraint (IR), et al.In the first chapter, the application of MRAT and the background of our research are introduced. Several related conceptions of Mechanism Design of MRAT have been introduced, the marketing patterns of E-Procurement are summarized. And then, the research purpose of this paper is pointed out based on the results of current research. In the second chapter, operation flow in E-Procurement is divided into five phases: searching phase, learning phase, order assigning, negotiation phase and implement phase, in each of which, principals and strategies that buyers should follow are pointed out. In the third chapter, the kernel problems of MRAT (cost estimation and order assigning) are mainly discussed. Based on Multi-attribution Decision Problem and Multi-attribution Utility theory, the estimation model of supplier's cost function is constructed with the assumption of Myopic best-response bids and undistorted bids, and the offer distribution optimization model is constructed with consideration of the constraints of procurement cost, procurement quantities and order period. In the forth chapter, based on the model constructed, with two non-price attribution (the quantity attribution and delivery time) considered simply, the instance of MRAT is given and is simulated.In this paper, based on the situation of the large quantity of procurement and the high level of standard, a set of complete mechanism of MRAT, with searching suppliers, cost function of suppliers learning, bid winner determination, offer distribution and offer signing. Rationality and feasibility in the flow of the procurement designed in this paper are analyzed with Game Theory. The problem of cost estimation is discussed in the condition of incomplete information, which helps buyers estimate supplier's cost function and assign order. The computation of the 0-1 knapsack problem is decreased, the computation time is shortened and the decision efficiency is improved because of the department of bid winner determination and order assigning, and combinatorial explosion is avoided availably in NP-hard Problem.