JP2005129078A - Electronic auction device and its system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an electronic auction device and an electronic auction system capable of efficiently allocating goods by solving problems related to an overload of a server, the congestion of a network, and information disclosure. <P>SOLUTION: This electronic auction device and this electronic auction system can be realized by providing them with a means for storing early fixed highest bidding amount data received from respective buyer devices in an automatic substitutional bidding management database after a bid-up processing part of the electronic auction device extracts a threshold value and privilege data from a commodity management database to transmit them to the buyer devices, and for transmitting information of presence/absence of the early fixed highest bidding amount to the buyer devices after a presence/absence assertion period elapses. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to an electronic auction apparatus that performs a bid auction for determining the allocation of a product and the sales price when the product is sold through a network.

  The most commonly used protocol for auctions on the Internet is a bid auction with an end time (British auction). The transaction procedure is as follows.

  First, a seller registers a product (good) on an auction site. The buyer can view the start time / end time of goods and auctions via the network. At the start time, the buyer can place a bid. The new bid must be greater than the current highest bid price. Usually, the minimum bid value is set. All bids that are larger than the current highest bid value are open to other bidders. When the end time comes, the bidder who is bidding at the highest price at that time becomes the winner, gets an allocation of goods, and pays the seller an amount equal to his bid price.

  Many auction sites on the Internet provide automatic proxy bidding agents (proxy agents). This is a system for bidding for the lowest price accepted on behalf of a person. All an individual bidder does is set the highest bid amount, the highest value that the bidder is willing to pay for the good. If another person places a bid that exceeds his bid price, and within the limits of the maximum bid price, the automatic proxy bid agent bids the bid price of the other person plus the minimum bid price. This eliminates the need for the buyer to sit in front of the calculator throughout the auction and monitor bids.

  When self is a successful bidder using an automatic proxy bidding agent, the payment amount is the highest bidding price, excluding self, plus the minimum bid price. Accordingly, the true evaluation value for the goods may be set as the highest bid amount as it is.

  This is because the amount of payment is substantially determined by the bid price of the other person, and the amount of payment cannot be reduced even if the maximum bid is set low. Also, if you do not receive the allocation of goods when you set the true evaluation value, if you set the maximum payment amount larger than your true evaluation value to receive the allocation of goods, the payment amount will be your own true value It is because it becomes more than evaluation value and loses in the transaction.

  The Internet auction usage guide also states that the maximum bid amount of an automatic proxy bidding agent should be set to the value that it is willing to pay. However, in actual auctions, there is a phenomenon in which individual bidders change the setting of the maximum bid amount many times, and a bid behavior is observed in which a significant part of the bid is bid immediately before the end time of the auction. Has been.

  This action may cause “a problem of server overload and network congestion” and “a problem of revealing information about goods”. When these problems arise, efficient allocation of goods cannot be achieved.

  In other words, if a concentration of bids immediately before the end time occurs, the auction server may fail to receive all bids, that is, there is a problem of “bid dropout”. Note that the problem caused by such a failure to receive a bid may be alleviated to some extent by distributing transactions using a plurality of servers.

  The other issue, information disclosure, is a problem related to the evaluation of goods. The evaluation of this good is usually more or less influenced by the evaluation of others. In particular, if the bidder is unable to judge the quality and authenticity of the goods sufficiently, it can be said that the evaluation is more susceptible to the evaluation of others.

  For example, in a painting auction, it is difficult to judge the authenticity of a picture unless the bidder is an expert in painting transactions. However, even if it is an ordinary buyer, if the other party actively bids for a large amount of money, the buyer may assume that the picture is real and bid according to the real picture. it can.

Sharing of evaluation information of goods by information disclosure is an advantage of an open auction over a sealed auction. Here, the sealed type refers to a format in which all bids are not released until the auction is completed. On the other hand, the open type refers to a format in which all bids are open to all bidders during an auction. Auctions are one of the open types.

  However, if almost no bid is placed until just before the bid end time, the bidder cannot obtain information on the goods through the bid action of the other person. This problem cannot be solved easily even if the functions of servers and network equipment are enhanced.

(Summary)
In the auction with the end time set as described above, the following problems occur due to concentration of bids immediately before the end time.

  (1) The problem that bids are dropped due to server overload and network congestion.

  (2) The problem that information about goods cannot be obtained due to failure to reveal information.

  Due to these problems, efficient allocation of goods cannot be achieved.

  Accordingly, an object of the present invention is to provide an electronic auction apparatus and an electronic auction system that can efficiently allocate goods by solving problems related to server overload, network congestion, and information disclosure. Is to provide.

  The electronic auction apparatus according to claim 1 is an electronic auction apparatus having an automatic proxy bidding agent function connected to a buyer apparatus via a communication network, and the electronic auction apparatus stores product information. A product management database, an automatic proxy bid management database storing maximum bid information, and a bid processing unit, wherein the bid processing unit extracts a threshold value and privilege data from the product management database and extracts the buyer After the transmission to the device, the early fixed maximum bid amount data received from each buyer device is stored in the automatic proxy bid management database, and information on the presence or absence of the early fixed maximum bid amount for the buyer device after the presence / absence expression period has passed. It is characterized by transmitting to.

  The electronic auction apparatus according to claim 2, wherein the auction processing unit extracts the current highest bid price data of the target product of the auction from the bid management database, and is larger in the automatic proxy bid management database. If there is maximum bid amount data, the process of transmitting new maximum bid amount data to the buyer device is continued until the end time of the auction, and the highest bid amount transmitted from the buyer device of the auction winner is determined after the end time. It is characterized by further providing.

  The electronic auction system according to claim 3 is an electronic auction system having an automatic proxy bidding agent function connected to a buyer device via a communication network, and the electronic auction device stores product information. A product management database, an automatic proxy bid management database storing maximum bid information, and a bid processing unit, wherein the bid processing unit extracts a threshold value and privilege data from the product management database and extracts the buyer After the transmission to the device, the early fixed maximum bid amount data received from each buyer device is stored in the automatic proxy bid management database, and information on the presence or absence of the early fixed maximum bid amount for the buyer device after the presence / absence expression period has passed. The buyer device receives the presence / absence information and changes the set maximum bid amount. And butterflies.

  According to the present invention, the electronic auction apparatus (auctioneer) is the highest by the automatic proxy bidding agent for the auction for determining the product allocation and the sales price when the product is auctioned through the network. For buyer devices that send thresholds and compensation data for early fixation of bid amount to buyer's device, set whether buyer device will fix early or not, and fix highest bid amount early , Because given benefits (privileges such as discounts on payments and compensation payments) were avoided, bid concentration just before the end time was avoided, server overload, network congestion, and information disclosure The problem can be mitigated and linked to an increase in social surplus (ie, the social preference of the allocation of goods), which enables the efficient allocation of goods. It was.

  The best mode for carrying out the present invention will be described below in detail with reference to the drawings.

[First example]
A first embodiment of the present invention will be described with reference to FIGS.

(Overview)
First, an outline of the present invention will be described.
In the auction protocol according to the present invention, an automatic proxy bid agent performs early fixing of the highest bid price (hereinafter, the highest bid price for which early fixing has been performed is referred to as an “early fixed highest bid price”). For bidders who allow and fix the maximum bid early, the following benefits:
・ If you are the winner of an auction, you can get a discount or pay compensation
・ Even if the auction loser, the benefit of being paid a compensation to the highest bidder with the highest fixed bid,
give.

  In this way, by giving a privilege to the bidder of the early fixed maximum bid amount, the bid concentration immediately before the end time can be eased.

(Concrete example)
Hereinafter, a specific example will be described.

<System configuration>
First, the configuration of a system applicable to an electronic auction according to the present invention will be described.

  FIG. 2 shows a configuration example of the electronic auction system 1.

  The system 1 includes an electronic auction apparatus 10 as a server side apparatus, and a seller apparatus 100 / buyer apparatus 110 as a client side apparatus. These devices are connected by a communication network 200.

(Electronic auction equipment)
The electronic auction apparatus 10 will be described.
The bidder management database (DB) 11 stores information such as contact information and past transaction history, which are information related to auction participants, that is, bidders.

  The merchandise management database (DB) 12 stores information related to merchandise that has been exhibited in the auction, that is, information such as the seller, the type of merchandise, the number, the start time, the end time, the threshold, and compensation money.

  The bid management database (DB) 13 stores products, history information on bidders and bid amounts for each product.

  The automatic proxy bid management database (DB) 14 stores information on products, bidders and maximum bids (including early fixed maximum bids).

  The bidder registration unit 15 has a function of registering / changing information such as contact information transmitted from a bidder who is new or who wants to change registered information in the bidder management database 11.

  The merchandise registration unit 16 has a function of registering information on desired merchandise items transmitted from the seller device in the merchandise management database 12.

  The product search unit 17 has a function of searching the product management database 12 and creating a product list of a type specified by the buyer device for transmission to the buyer device.

  The automatic proxy bid processing unit 18 has a function of registering information related to the product and the highest bid amount transmitted by the buyer device in the automatic proxy bid management database 14.

The bid-up processing unit 19 has a function of executing the following processes (1) to (4).
(1) Obtain the product and the current maximum bid data for the product from the bid management database 13.
(2) The automatic proxy bid database 14 is searched to check whether or not there is a setting of maximum bid price data larger than the current maximum bid price data for the product.
(3) If the highest bid amount data is set in (2), bid management is performed together with the bidder's information as the new highest bid amount data obtained by adding the minimum bid amount data to the current highest bid amount data. Register in the database.
(4) If there is no larger maximum bid data set, do nothing.

  The auction processing unit 19 repeats such processing until the bid end time for each product comes. When the end time comes, the successful bid processing unit 20 is notified.

  The successful bid processing unit 20 obtains the highest bid price from the bid management database 13, uses this bid price as the payment amount, obtains contact information of the seller / buyer from the bidder management database 11, and obtains the successful bid result as the seller device and Send to buyer device.

  The control unit 21 performs overall control of the entire system and control of processing as shown in the flowchart of FIG. 1 according to the present invention.

  The communication unit 22 transmits and receives information to and from the seller device 100 and the buyer device 110.

(Seller / buyer)
The seller apparatus 100 and the buyer apparatus 110 will be described.

  In this example, a single good, single unit auction is handled. Assume that there are two types of buyers. One type of buyer can see the quality and authenticity by looking at the goods. The other type of buyer cannot see the quality and authenticity of the goods. The former is called an expert buyer and the latter is called an ordinary buyer.

  It is also assumed that there are two types of quality of goods, high quality and low quality. Assume that individual buyer evaluations of goods are linked to other buyer evaluations. In the present invention, the evaluation is linked means that if an expert buyer has a higher evaluation value, the evaluation value of an ordinary buyer is likely to be higher than low. That is, if an ordinary buyer knows the fact that an expert buyer has a high reputation for the goods, the ordinary buyer will assume that the goods are of high quality and that the goods give. However, for both professional buyers and ordinary buyers, once the quality of the goods becomes clear, the evaluation value of each individual buyer is independent of the evaluation value of others.

  Assume that the evaluation value of each buyer for a good is obtained from a probability distribution. This probability distribution is assumed to be known. If the goods are of high quality, the value of the goods is the interval

According to the probability density function f (cumulative distribution function F) above, if the goods are of low quality, the evaluation value of the goods is

It is assumed that the probability density function g (cumulative distribution function G) above is followed. In this article, to simplify the discussion,

Assuming

  The individual buyer's utility is defined as the difference between the true evaluation value and the payment amount if the buyer wins the auction, and is defined as 0 if the buyer loses the auction. However, if the payment is greater than the true value, the buyer will have a very large negative utility. That is, it is assumed that each buyer is risk-neutral when the difference between the evaluation value and the payment amount is positive, but is risk-loving when the difference is negative.

  Here, risk-neutral means that the behavior is determined in consideration of the magnitude of the expected value. Risk-loving means that if the amount of loss increases from 0 even a little, the decrease in utility becomes very large, but if the amount of loss increases very much, even if the amount increases further, the utility This means that the degree of decrease is small.

  Also assume that there are n expert buyers and m ordinary buyers.

Represents a list of expert buyer ratings,

Let denote a list of evaluation values for ordinary buyers. For high quality goods, the value q takes 1 and for low quality goods, the value q takes 2. further,

と

When

Is a list of each evaluation value in descending order.

は

Is

The k-th highest value in

は

Is

It is assumed that the kth highest value is indicated.

<System operation>
Next, the operation of this system will be described.

  In this example, a bid auction protocol (British auction protocol) with a bid end time set is considered.

  In this protocol, the new bid must be greater than the current highest bid value. Usually, a minimum bid value ε is set. All bids are immediately published to other bidders. When the end time comes, the buyer who is bidding at the highest price at that time becomes the winner and gets the goods. The amount to be paid is equal to the bid price (the second highest bid price + ε as a whole).

  In this example, it is assumed that an automatic proxy bidding agent (proxy agent) is used.

  This automatic proxy bidding agent is a system for bidding for the lowest price accepted on behalf of a person. What an individual bidder does is set only the highest bid amount, that is, the highest amount that the bidder is willing to pay for the good.

Here, in order to handle the auction with a bid end time setting, an expression of time is introduced. Assuming a hypothetical time t, the auction ends at time t = t _end . The end time is guided before the start of the auction and is not changed.

  Also, in this example, it is assumed that the auction server can accept a sufficient number of bids during time 1 for bids from the automatic proxy bid agent. However, it is assumed that time 1 is required for each bidder to change the maximum bid setting of the automatic proxy bid agent.

  In this example, it is assumed that individual buyers are symmetric. This can be formulated as individual buyers observe a private signal that determines their type. Therefore, in the following discussion, we will examine symmetric equilibrium. There is no collusion between buyers. The buyer means a bidder, and it is natural that the bidder and the bidder evaluate the goods, but the bidder (buyer) among the bidders (buyer) among the content described below. The process executed means a process executed by the bidder between the electronic auction devices that are bidders using the buyer device, and the information processing performed by the bidders means that the bidders use the electronic auction devices. It means a process executed by the electronic auction apparatus.

  In the following, each example of an auction will be described and explained.

<Definition 1 Strategy A:>
Individual expert buyers observe the goods and, if the goods are of low quality, a true value for the low quality goods

Is set to the highest bid of the automated proxy bid agent, and if the goods are of high quality, the true evaluation value for the high quality goods

To the highest bid for an automated proxy bid agent.

  Individual ordinary buyers will begin with a true valuation for low-quality goods

To the highest bid for an automated proxy bid agent. Then, by looking at the v ¹ or more of the bid, the upper limit of the automatic proxy bidding agent

To true value for high quality goods

Change to

(Auction example 1: Auction without a bid end time)
As an auction example 1, a bid auction without setting a bid end time will be described.

Proposition 1) When a bid-up auction protocol without setting a bid end time is used, that is, when t _end = ∞, (n + m) sets of strategies A are balanced.

  Here, “balanced strategy” means that when other bidders ((n + m−1)) follow the strategy, it is best that the self also follow the strategy. It can be considered that the states that appear in actual auctions are balanced.

  Proof 1) First, consider an expert buyer.

  Expert buyers cannot control the number of bidders, even if they delay expensive bidding. Therefore, since the evaluation value of another person's goods does not change for an expert buyer, if the maximum bid amount of the automatic proxy bidding agent is set, it can be set immediately.

  Now suppose that the expert buyer sets his true valuation value to the automatic proxy agent's highest bid. At this time, if the buyer becomes the winner of the auction, even if the buyer sets the maximum bid amount to a different value, the buyer cannot obtain a larger utility. This is because the payment amount is the sum of the second largest bid price and the minimum bid price, and is substantially determined by the bid price of the other person (second largest bid price). On the other hand, if the buyer loses the auction, the buyer will not be able to gain greater utility even if the highest bid is set to a different value. This is because even if a higher maximum bid is set and the auction is won, the amount paid becomes larger than the self-evaluation value. Therefore, Strategy A is the optimal reaction for the expert buyer.

  Next, consider an ordinary buyer.

  If the average buyer knows the quality of the goods, it is optimal to set the true evaluation value corresponding to the quality as the maximum bid price of the automatic proxy bid agent, as in the case of the expert buyer's argument.

Here, when the other buyer to follow the strategy A, v ¹ to carry out the above bid is a buyer of the experts. Since the buyer of the experts it is possible to know the quality of goods, if you look at the bid ordinary buyer is higher than v ^1, the buyer can the goods is concluded that a high quality.

  Here, before a normal buyer knows the quality of the goods, the true evaluation value for the low-quality goods

Suppose you set a higher value for the highest bid. At this time, if it turns out that the goods are of low quality, the payment amount of the ordinary buyer may be larger than his own evaluation value. So if the quality of the goods is unknown, the average buyer

Do not set a higher value for the highest bid. Therefore, Strategy A is an optimal reaction for ordinary buyers.
(Auction example 2: Bidding auction with bid end time set)
Next, as an auction example 2, a bid auction with a bid end time set will be described.

Proposition 2) When a bidding auction protocol with a bid end time set is used and the goods are of high quality, the expert buyer will consider the self for high quality goods at time t _end -1 immediately before the end time. True evaluation value of

It is best to set the maximum bid for an automated proxy bid agent. In this case, a normal buyer cannot win the auction.

  Proof 2) Expert buyer time

In addition, if the true evaluation value of the self is set to the highest bid amount of the automatic proxy bidding agent, the bid value exceeds v ¹ when there are two or more specialist buyers. This means conveying information to the ordinary buyer that the goods are of high quality. Therefore, the auction is between n + m bidders.

On the other hand, if the buyer of an individual expert sets his true evaluation value at the maximum bid amount of the automatic proxy bid agent at time t = t _end −1, an ordinary buyer may place a bid larger than v ^1. Can not. This is because it takes time 1 to change the setting value of the maximum bid amount of the automatic proxy bidding agent.

  Here, before knowing the quality of the goods, the average value for a low-quality goods by ordinary buyers

Note that the larger the bid, the buyer may get a negative profit. Therefore, in this case, the number of bidders is substantially n persons who are only expert buyers. A substantial reduction in the number of bidders will benefit professional buyers. In addition, since it is assumed that all bids from the automatic proxy bid agent will be accepted, even if the highest bid is set at time t = t _end −1, the expert buyer suffers from this. There is no. Therefore, the proposition is established.

  Here, the load on the server will be described. The number of bids that the server will accept just before the bid end time is that if the goods are of high quality,

It becomes.
(Auction example 3: Auction with fixed maximum bid)
Next, as auction example 3, an auction with a fixed maximum bid amount will be described.

  When a bid auction protocol with a bid end time set is used, information about the quality of the goods is not communicated to ordinary buyers during the auction, even though it is an open auction. This raises the problem that when the number of expert buyers is less than the number of ordinary buyers, efficient allocation of goods cannot be achieved or the seller's revenue does not increase.

  One solution to this problem is for the bidder or seller to guarantee the quality of the goods to the buyer. In a general cooperative value model, it is dominant for competitors or sellers to honestly guide the quality of goods. That is, when the goods are of high quality, the goods are expected to be high quality, and even if the goods are of low quality, the information is not concealed and the seller's expected income is shown by guiding the goods to be of low quality. Is supposed to increase. However, in an auction between consumers, the seller is not necessarily an expert. Even if you are an expert, making a specific arrangement for quality assurance is a cumbersome task.

  In this example, a new protocol is proposed to solve the problem caused by concentration of bids immediately before the bid end time. The basic idea is that the buyer's device that fixed the maximum bid amount of the automatic proxy bid agent at the start of the auction will receive payment discount data and compensation data when it wins the auction, If such a buyer's device fails to obtain goods, compensation data is transmitted. This eliminates the need for the buyer's device to check the bid data of the other person and provides an incentive to make a high bid at an early stage.

(Auction Case 3: Flowchart)
This auction example 3 is demonstrated based on a flowchart.

  FIG. 1 shows the process of a bid auction with a fixed maximum bid amount. The normal process of this auction is a process performed by the auction processing unit 19 as described in paragraph 41. The auction processing with the highest bid amount fixed is also performed by the auction processing unit 19 in the same manner.

  In step S1, the electronic auction apparatus transmits the auction start time and end time to the buyer apparatus.

In step S2, and transmits the compensation p _c Amount data with the threshold value v _t the buyer device for.

In step S3, an early fixation of the highest bid amount is made based on the condition that the fixed value must be _{equal to} or greater than the threshold value vt for early fixation of the highest bid amount during the early assertion period of the highest bid amount. Receive the desired maximum bid data from each buyer device.

  In step S4, after the end of the presence / absence announcement period, the electronic auction apparatus, if the highest bid data desired to be fixed early is transmitted from two or more of the buyer apparatuses, the information indicating that the maximum bid amount is fixed at an early stage. In other cases, information on the highest bid amount without early fixing is transmitted.

  In step S5, after the information is transmitted, a normal auction is performed in which only a bid amount larger than the current highest bid amount is accepted (see paragraph 42 for the processing procedure of the auction).

  In step S6, when the bid end time comes, the highest buyer device is assigned as the auction winner, the goods are allocated, and the bid amount data of that person is used as the payment amount data. The processes from step S1 to step S6 are executed by the auction processing unit 19.

  In step S7, based on the data in the automatic proxy bid management database, it is checked whether or not "the highest bid price data is fixed early", that is, whether or not there is an early highest bid price data. If fixed, the process proceeds to step S8.

  In step S8, a check is made as to whether or not the buyer device is the winner A of the auction that has made the highest bid among the buyer devices with the highest bid amount fixed at an early stage. In the case of the buyer device of the auction winner A, the process proceeds to step S9, and in the case of the buyer device of the auction loser, the process proceeds to step S12.

  In step S9, it is checked whether or not the buyer B of the person B who made the second highest bid in the auction also fixes the highest bid amount. If the buyer's device of person B who performed the second highest bidder device also fixes the highest bid amount, the process proceeds to step S10, and if not fixed, the process proceeds to step S11.

In step S10, the buyer B of the person B who placed the second highest bid also fixes the highest bid amount, and therefore no compensation _pc is paid to anyone.

In step S11, the buyer apparatus of person B, which made the high bid in the second is, because it does not secure the highest bid, the auction buyer apparatus of the winner A of which was the highest bid, pay compensation p _c.

In the step S12, in the auction of the loser of the buyer device in the fixing of the highest bid, the bid amount is for a maximum of persons of the buyer equipment, pay compensation p _c. As explained in paragraph 44 above, the highest bid amount is obtained and this bid amount is used as the payment amount, the seller / buyer contact information is obtained from the bidder management database 11, and the successful bid result is sent to the seller device and the buyer device The transmission process is a process performed by the successful bid processing unit 20. Similarly, the successful bid processing unit 20 also processes the successful bid result of the auction with the highest bid amount fixed.

  As can be seen from the above processing, the automatic proxy bidding agent allowed the buyer device to fix the highest bid amount early, and in the buyer device that fixed the highest bid amount early, In addition to paying discounts and paying compensation, the auction loser is given the privilege of paying compensation to the person with the highest maximum bid setting. Can be relaxed.

(Auction example 3: specific example)
The specific example of this auction example 3 is given and the characteristic of the auction with fixed maximum bid amount is demonstrated. This example is an extension of the auction protocol with a bid end time set.

a) The competitor sets a threshold value v _t based on knowledge about the goods. When the buyer fixes the automatic bidding agent's maximum bid amount, the maximum bid amount must be greater than or _{equal to the} threshold vt. The threshold value may be set to any threshold value as long as the quality of the goods can be sufficiently distinguished. In this example, v _t = v ¹ is set. The bidder guides this threshold before starting the auction.

b) auction who sets the compensation _{p c.} This value is

Must be a value that satisfies The meaning of these values is explained in the proof of Proposition 3 below. The bidder informs the compensation value before the auction starts.

  c) At the start of the auction, a period for each buyer to inform the bidder of the intention to fix the maximum bid amount of the automatic proxy bid agent is set. This period need not be so long. For example, if a bid period is 7 days for a certain product, it is sufficient to set the intention expression period to 1 day. Even during this period, normal bidding is accepted.

d) The bidder adds to the set S _ixf a buyer's device that fixes the maximum bid amount of the automated proxy bid agent. The initial value of S _fix is an empty set.

  ・ After the period for expressing the maximum bid amount has passed, the bidder informs the buyer of the highest bid amount. This is called signal Q. If there are two or more buyers who wish to fix the highest bid, the bidder informs them that the highest bid is fixed (Q = 1). In other cases, the bidder informs that there is no fixed maximum bid (Q = 0). If only one buyer wishes to fix the maximum bid, the maximum bid is not fixed. The maximum bid value itself and the number of buyer devices with a fixed maximum bid amount are not provided.

  In the strategy B described below, the guidance that the highest bid amount is fixed means that the goods are of high quality.

  e) After a period of intention to fix the maximum bid amount, a normal auction will be held until the bid end time.

f) When the bidding end time t = t _end comes, the auction ends. At this point, the buyer with the highest bid becomes the winner, receives the allocation of goods, and pays an amount equal to his bid (second highest bid + minimum bid ε).

When there are a plurality of buyers having the highest evaluation value, the buyer who has received the highest bid first reaches the auction server is the winner. Usually, in order to maintain fairness, processing such as giving priority to a bid from a specific seller is not performed, so that the winner is virtually randomly selected from buyers having the highest evaluation value. The payment amount is the bid amount, that is, the highest evaluation value itself. (The second highest bid value is the highest evaluation value-the minimum bid value ε.)
g) If the auction who has guided that there is a fixed maximum bid (Q = 1), the seller pays the compensation p _c to the buyer, which was the highest bid in the buyer included in the set S _fix. However, if both the auction winner and the buyer device that made the second highest bid are included in the set S _fix , no compensation will be paid to anyone.

Here, the handling when there are a plurality of buyers who have fixed the highest bid price in the set S _fix will be described. In the present embodiment, as in the above-described winner determination, the buyer whose highest bid (in S _fix ) is accepted by the auction server is entitled to obtain compensation. Usually, in order to maintain fairness, priority is not given to bidding from a specific seller, so in effect, buyers who can obtain compensation are among the buyers who set the highest bid price. Will be chosen at random.

As another method, a method of using a time at which the highest bid price is fixed is also conceivable. That is, when the highest bid price is fixed, the time information is also transmitted to the auction server. Then, among the bids from the buyers included in the set S _fix , priority is given to the bid of the buyer who sets the highest value at the earliest time. In this case, among the buyers who set the highest bid price fixed, the buyer who set the earliest time gets the right to obtain compensation.

<Definition 2 Strategy B:>
(1) If the goods are of high quality, the expert buyer i is the true value for the quality goods

Is set to the highest bid of the automatic proxy bid agent, and the highest bid is fixed. Also, if the goods are of low quality, the expert buyer will give a true rating for the low quality goods

To the highest bid for an automated proxy bid agent.

  (2) Ordinary buyer j starts with a true valuation for low quality goods

To the maximum bid of the automated proxy bid agent. After that, if you see signal Q = 1 with fixed maximum bid, you can see the maximum bid of the automated proxy bid agent.

True value for high quality goods

Change to If you see signal Q = 0 with no fixed maximum bid amount, the maximum bid amount of the automatic proxy bid agent is left as it is.

Keep it.

(Auction Case 4: Auction Auction)
Next, as an auction example 4, a bid auction will be described.

  Proposition 3) (n + m) sets of strategy B are balanced.

  Proof 3) First, consider an expert buyer.

Assume that the goods are of high quality. If expert buyer i does not fix the maximum bid amount of the automated proxy bid agent, from Proposition 2, the true evaluation value for the quality goods at time t _end −1

The best response is to set to the highest bid. At this time, it is substantially an auction among n buyers. In this case, the expected utility is expressed by equation (1).

In equation (1), n−1 | represents the number of buyer candidates who have the second highest bid. In order for the second highest bid value to be z, the evaluation value of the other n-2 buyers must be smaller than z. This probability can be expressed as (F (z)) ⁿ⁻² . In this case, the payment amount of the expert buyer i is z + ε. Here, assuming that the minimum bid value ε is a minute amount, it will be ignored in the following discussion. Therefore, the utility of expert buyer i is

Is calculated. The probability that a buyer with the second highest bid has an evaluation value of z is f (z). From these discussions, equation (1) is obtained.

  On the other hand, if expert buyer i fixes the maximum bid of an automated proxy bid agent, the maximum bid is still

Setting to is the optimal reaction. The reason for this will be described below.

Suppose that expert buyer i sets a value greater than the true evaluation value. At this time, if the expert buyer i wins the auction and the buyer who has made the second highest bid is included in the set S _fix , the expert buyer i cannot obtain compensation. Therefore, in this case, the payment amount of the professional buyer i exceeds the true evaluation value. Here, since the buyer assumes that the difference between the valuation value and the payment amount is negative, the buyer has a very large negative utility.

A larger value is never set. Further, since the payment amount is determined by the bid price of the other person, even if a value smaller than the self evaluation value is set as the maximum bid amount, the payment amount cannot be reduced.

  Therefore, when the expert buyer i fixes the maximum bid amount of the automatic proxy bid agent, the expected utility is expressed by the equation (2). Here, the first term represents the utility obtained by winning the auction and obtaining the goods. The second term represents the utility of obtaining compensation money. The third term is a term corresponding to the case where compensation is not paid due to the exclusion rule.

Compensation _pc is

It is set to satisfy. Thus, if the goods are of high quality, the automatic bidding agent's maximum bid is

It is an optimal reaction for the professional buyer i.

  Next, consider the case where the goods are of low quality.

  If expert buyer i sets a value higher than the true valuation value for the automatic proxy bid agent's highest bid, the buyer may get a negative utility, as in the case of high quality. Therefore, the expert buyer does not set a value higher than his evaluation value as the highest bid. In this case, since the true evaluation value for the goods is smaller than the threshold value, the maximum bid amount cannot be fixed. Further, when winning the auction by setting its own true evaluation value as the highest bid amount, the payment amount does not decrease even if it is set smaller than the true evaluation value of the highest bid amount. Thus, if the goods are of low quality, the maximum bid of the automated proxy bid agent is

Is an optimal reaction for the expert buyer i.

  Next, consider an ordinary buyer.

  Suppose that an ordinary buyer fixes the automatic bidding agent's maximum bid before looking at the signal Q indicating whether the maximum bid is fixed or not. In this case, if the goods are of low quality and the buyer wins the auction, the buyer's payment may exceed his true value.

  Also, if the goods are of low quality, the expert buyer does not fix the maximum bid of the automated proxy bid agent. Therefore, since there is only one buyer who wishes to fix the highest bid amount, the signal Q = 1 indicating that the highest bid amount is fixed is not transmitted. In this case, no compensation is paid to the ordinary buyer.

  Therefore, since buyers are assumed to be risk enthusiastic about the negative difference between the valuation value and the payment amount, ordinary buyers do not take action to fix the maximum bid amount of the automatic proxy bid agent. Absent.

  If the ordinary buyer does not take action to fix the highest bid, the same as the proof of Proposition 1, the ordinary buyer can set his true evaluation value to the highest bid of the automatic proxy bid agent. It can be proved that the reaction is optimal.

  Proposition 3 implicitly assumes that the number of expert buyers and the number of ordinary buyers are common knowledge. This assumption may seem too strong. However, not all expert buyers need to take action to fix the maximum bid of an automated proxy bid agent. In terms of communicating information about the quality of goods to ordinary buyers, it is sufficient for two or more specialist buyers to fix the maximum bid. Therefore, it is considered that in many cases, information disclosure can be achieved even if the amount of compensation is set using the values of n and m predicted by the competitor.

Compared to a bid auction protocol with a bid end time set, when using the bid auction protocol according to the present invention, it can be expected that an expensive bid will appear earlier. This is because the expert buyer locks the maximum bid of the automated proxy bid agent. As a result, the problem of server overload and network congestion immediately before the bid end time can be alleviated.
(Compensation calculation example)
Next, a calculation example of compensation money will be described.

Example 1) Uniform distribution

Assuming By using the partial integration method, we obtain:

  Is calculated in the same way,

As the value _{p c} satisfying, obtain the following.

Here, p ′ _c is a value when the equal sign holds in the above equation,

In

It shall be

In between

Is an increasing function with respect to

Is all

Against

Meet.

Is obtained as follows.

Therefore, as the value of p _c of bid up auction protocol according to the present invention, if using the p _"c, strategy B is balanced.

(Evaluation of bidding auction protocol)
Next, evaluation of the performance of the auction auction protocol in this system will be described.

  In this chapter, the performance of the auction protocol according to the present invention is evaluated by computer simulation.

  Here, from the viewpoint of social surplus and seller's income, the auction auction protocol according to the present invention is compared with a normal auction protocol with a bid end time set. Here, the social surplus is calculated as the sum of the utility of the buyer device of all bidders including the seller, and represents the social desirability of the allocation of goods.

  When the goods are of low quality, the normal auction protocol with a bid end time set and the auction protocol according to the present invention give the same result. Therefore, it is sufficient to examine only when the goods are of high quality. The experimental setup is shown below.

  The evaluation value of each buyer for high quality goods is uniformly distributed on the interval [1000, 2000]

And more. The total number of buyer devices is 100. In this article, the ratio of expert buyers is varied from 0.1 to 0.9. A ratio of 0.1 means that there are 10 expert buyers (100 × 0.1 = 10) and 90 ordinary buyers (100−10 = 90). In addition, it is assumed that the evaluation value of the seller's goods is zero. This means that the utility of the seller is the difference between the selling price of the goods and the amount of compensation paid. The threshold value v _t is set to 1000, and the compensation money value is expressed by the equation (3).

Set to. For each case, 100 examples were generated and the average value was calculated.

  FIG. 3 shows a comparison result regarding social surplus.

  The horizontal axis represents the number of buyer devices of expert buyers among the buyer devices of 100 bidders. The vertical axis represents the average value of social surplus obtained from 100 cases. A is an example of a bid auction protocol according to the present invention. B is an example according to a conventional protocol.

  FIG. 4 shows a comparison result on seller's income.

  The horizontal axis represents the number of buyer devices of expert buyers among the buyer devices of 100 bidders. The vertical axis represents the average value of seller's income obtained from 100 cases. C is an example of a bid auction protocol according to the present invention. D is an example according to a conventional protocol.

  3 and 4 show the following (1) and (2).

  (1) The auction protocol method according to the present invention increases social surplus.

  Social surplus is determined by the true value of the auction winner for the goods. Therefore, if the number of buyer devices increases and a buyer with a higher evaluation value appears, the social surplus increases.

  FIG. 3 shows that the increase in social surplus increases when the ratio of expert buyers is small. This is because in this case, the fixed maximum bid amount of the automatic proxy bid agent will cause many ordinary buyers to participate in the auction of high quality goods.

  (2) When using the auction protocol method according to the present invention, if the ratio of specialist buyers is small, the increase in sellers' income will increase, but as the percentage of specialist buyers increases, The increase in income is small.

  Intuitive explanation is as follows.

  The auction protocol method according to the present invention is based on the condition that the same level of utility is given to the buyer of the expert. It can be seen that the goods are handed over to the buyer with the highest evaluation value. Thus, if the second highest evaluation value among all buyers is sufficiently large compared to the second highest evaluation value among experts, the seller will increase the revenue at the auction exceeding the payment of compensation. Is obtained. However, if the second highest valuation value among experts is sufficiently large and that value is close to the second highest valuation value among all buyers, the increase in revenue at the auction will be small. In this case, the increase in seller's revenue is offset by the payment of compensation.

  In auctions between consumers, the ratio of expert buyers can often be expected not to be so large. Thus, from the experimental results obtained, in many cases, the inventive device is considered advantageous to the seller.

[Second example]
Next, a second embodiment of the present invention will be described. The description of the same parts as those in the first example described above is omitted, and the same reference numerals are given.

  In the first example described above, it was assumed that the buyer would get a large negative utility if the payment amount would be greater than the true valuation value. In other words, it was assumed that the buyer would not choose such an action if he could lose.

  However, for example, if the expected value of earned profit is large, such as a company, not an individual trader, it may be appropriate to assume a form of utility that allows a small loss.

  In this case, strategy B is not balanced in the inventive protocol described in the previous embodiment. For low-quality goods,

For the average buyer with a close evaluation value, before looking v ¹ or more of the bid, the better to a bid for the goods of high quality (v ¹ or more of the bid), there is a case of expected utility increases .

  However, when the following changes are made to the auction protocol for the auction according to the present invention, the evaluation value varies greatly depending on the quality of the goods.

If it is much larger), strategy B is balanced. This means that this protocol can be used to achieve efficient allocation of goods and increase sellers' revenue.

  The following two points are changed with respect to the first example.

Changes 1) Place a dummy buyer. As a buyer, set the highest bid value for automatic proxy bidding agent to v ^1, carry out the bid. If the dummy buyer becomes the winner, the dummy buyer withdraws the bid and the buyer with the second highest bid becomes the winner. The payment amount is the bid price (the third highest bid price + ε). Bid withdrawal is only allowed for dummy agents.

Among the agents that have fixed the highest bid value of the change point 2) Automatic proxy bidding agent, the buyer device in the bid of the highest value, always pay compensation p _c.

  From change 1

If the value is much larger, that is, if the value of the goods differs greatly depending on whether the goods are of high quality or low quality, the utility of the buyer is always expressed by the difference between the true value of the goods and the payment amount. Even in this case, it is the best response for ordinary buyers to follow strategy B.

  Also, from change 2, even if the buyer's utility is always expressed as the difference between the true value of the goods and the amount paid, the expert buyer will see the true value as the highest bid for the automated proxy bid agent. Setting is the optimal response.

  It is shown by the following that setting a true evaluation value is optimal for an expert buyer. Since the buyer is symmetric, the true evaluation value

A function b _i = which determines the highest bid value b _i according to

think of. At this time, similarly to Expression (2), the expected utility of the buyer of the expert when the bid price is fixed is expressed by Expression (6).

^{Here, B -1 (B (z)} ) = Note that become is z, by differentiating Equation (6) _{b i,}

It can be seen that the expected utility is maximized.

It is a flowchart which shows the auction with fixed highest bid amount which is the 1st and 2nd embodiment of this invention. 1 is a block diagram showing a configuration of an electronic auction system according to the present invention. It is explanatory drawing which shows a social surplus compared with the past. It is explanatory drawing which shows a seller's income compared with the past.

Explanation of symbols

1 Electronic Auction System 10 Electronic Auction Device 11 Bidder Management Database (DB)
12 Product management database (DB)
13 Bid management database (DB)
14 Automatic proxy bid management database (DB)
DESCRIPTION OF SYMBOLS 15 Bidder registration part 16 Product registration part 17 Product search part 18 Automatic proxy bid process part 19 Auction process part 20 Successful bid process part
100 Seller device 110 Buyer device 200 Communication network

Claims (3)

An electronic auction apparatus having an automatic proxy bidding agent function connected to a buyer apparatus via a communication network,
The electronic auction apparatus includes a product management database that stores product information;
An automated proxy bid management database to store maximum bid information;
With a bid processing section,
The auction processing unit extracts threshold value and privilege data from the product management database and transmits them to the buyer device, and then stores the early fixed maximum bid amount data received from each buyer device in the automatic proxy bid management database. An electronic auction apparatus that transmits information about the presence or absence of the early fixed maximum bid amount to the buyer apparatus after the presence / absence expression period has elapsed.   The auction processing unit extracts the current highest bid price data of the target product of the auction from the bid management database, and if there is higher highest bid price data in the automatic proxy bid management database, a new highest bid price data is obtained. The method further comprises: continuing the process of transmitting bid amount data to the buyer device until the auction end time, and determining the highest bid amount transmitted from the buyer device of the auction winner after the end time. Item 2. The electronic auction apparatus according to Item 1. An electronic auction system connected via a communication network to a buyer device having an automatic proxy bidding agent function,
The electronic auction apparatus includes a product management database that stores product information;
An automated proxy bid management database to store maximum bid information;
With a bid processing section,
The auction processing unit extracts threshold value and privilege data from the product management database and transmits them to the buyer device, and then stores the early fixed maximum bid amount data received from each buyer device in the automatic proxy bid management database. And sending information on the presence or absence of the early fixed maximum bid amount to the buyer device after the presence / absence declaration period has passed,
The buyer apparatus receives the presence / absence information and changes the set maximum bid amount.

Images