Research And Analysis On Bid-evaluation Behaviors Of Multi-attribute Reverse Auctions For Online Procurement

As the rapid development of the network information and electronic commerce, reverse auctions with multiple attributes are widely used for centralized procurements of large enterprises and government departments. This new purchasing method can not only overcome the corruption phenomenon during procurements, truly expand development space for the source, promote full and effective competition in the market, decrease the cost of purchase, realize the effectively configuration of market tendering resources, so it can bring huge economic benefits. At the same time it could provide a fair trading environment for suppliers. Therefore, as an important research field of electronic commerce, researchers would pay more and more attention to online reverse auction theory and method.What bid evaluation mechanisms are deployed become key factors in the fairness of auctions and resource allocation. It is the dominant factors of supply and demand in the supply chain. Bid evaluation for large construction projects or equipment procurements is work of a professional nature. A professional bid evaluation referee is called in as an expert. In practice, the attributes of auctioned objects can be divided into two types: technical and business. They are usually reviewed and scored by the latter in tandem fashion. This bid evaluation mechanism is referred to as a Grouped Multi-Attribute Auction (GMAA). Since the two groups of bid evaluation experts have difference objectives and responsibilities in this enterprise, they may be either cooperative or antagonistic during the bid evaluation process. In any case, their behaviors will certainly affect the fairness of the auction results.In view of the potential problems of the current evaluation mechanism, this paper is dedicated to the analysis and research of bidding evaluation behaviors in GMAA, achieve the following two purpose.(1) Try to demonstrate the several pressing problems within the existing bid evaluation processes that involve two groups of referee experts with scientific methods and means. On that basis, improve the bidding evaluation mechanism to to ensure fair evaluation.(2) Analyse the influence factors to decision-making behavior about bid evaluation experts, and we will enable those involved to choose from tested management practices and thereby channel experts'behaviors onto more reliable paths. In this way, they will can improve the fairness of, and help to optimize, the bidding evaluation process in general.Research contents and innovation points of this dissertation are presented as follows:(1) Aiming at the actual bid and tender status, by observing actual reverse auctions for procurements in several Chinese enterprises, we propose the Grouped Multi-Attribute Auction and outline the flow of grouped bidding evaluation. When human factors are taken into consideration, antagonisms and noncooperation between the interested parties seem inevitable. The necessity of the study is demonstrated.(2) Aiming at the current bidding evaluation mechanism, with the bid assessment data of a large tendering co., LTD. the for case, and then take example by simulation and behavioral analysis methods of experimental economics, we get the conclusion that the accumulation of antagonism exists surely through the analysis of experimental data. At last, we suggest:To set a group of integrative experts and some criteria related with two groups of attributes, and open both technical and business criteria to opposing groups. Based on the experimental data, the comparison and analysis about the three mechanisms are given.(3) Aiming at the long term evolution of bidding evaluation behaviors, this research focuses on the bid evaluation behaviors of referee experts employing an evolutionary game approach. It constructs a noncooperative static game model based on grouped bid evaluation behaviors in order to analyze the conditions necessary for achieving a long-term cooperative equilibrium between two groups. This construction refers, of course, to the pure strategy Nash equilibrium. It analyzes the evolutionary paths and factors influencing bid evaluation behaviors using the concept of the evolutionary game. We describe the effect of individuals'choice of policy on group behaviors. Similarly, a game payoff matrix between experts and administrative department is constructed to analyze the conditions how to achieve long term cooperation. To maximize clarity on these bid evaluation behaviors, we developed a simulation system based on the evolution process and ran it several times on a MATLAB GUI platform, which is convenient for adjusting relevant parameters. The simulation results make explicit a wider gamut of decision-making parameters as well as start-up variants on initial conditions that determine further (possible) evolutions.(4) Taking the bounded rationality of the bidding evaluation experts into consideration, some existing problems about the current mechanism are analyzed, and the relational graph of bidding evaluation behaviors about bidding experts is presented. Meanwhile, the qualitative model, which could influence the decision-making behavior of bidding evaluation experts, is established. The antagonistic feelings of two groups are studied by means of qualitative simulation (QSIM) and then the decision variable of different moments could be adjusted. Some significant results about the behavioral analysis have been obtained. And the simulation results indicated the effectiveness of this method. Because of the different benefit and responsibility, it is easy to bring about the accumulation of antagonistic feelings. As the difference increases, objectiveness becomes worse and worse.(5) Aiming at the complexity of the bid evaluation behavior, to verify the discovered points, a multi-agent-based approach is used to simulate the bid evaluation behaviors. The prototype system of a current bid evaluation mechanism is described by a multi-agent model. In the model, the behaviors of both auctioneer and bidders are described by the attributes of each agent and decision rules. Their interactions are described by the collaboration and information exchange between different agents. The procedure of multi-agent simulation is specifically presented. Then, the multi-agent simulation system is developed and repeatedly runs on the platform of Matlab GUI while adjusting relevant parameters. The results demonstrate that the accumulation of antagonistic feelings surely exist and the effectiveness of the modified bid evaluation mechanisms.