Method and system for setting an optimal preference policy for an auction

What is claimed is:

1. A computer implemented method for determining a preference policy for an auction to be conducted, said method comprising: selecting characteristics of said auction to be conducted; selecting a relevant bidding model for said auction to be conducted; estimating a structure of said auction to be conducted, said estimating comprises expressing unobservable variables in terms of observable bids, wherein said unobservable variables are expressed in terms of observable bids by inverting said bid model; predicting a bidding behavior for said auction to be conducted; predicting a first outcome of said auction to be conducted; and evaluating said first outcome of said auction to be conducted, wherein said evaluating comprises: selecting an optimal preference policy from a plurality of candidate preference policies for treating different groups of bidders differently, wherein said optimal preference policy comprises the candidate preference policy within a plurality having the highest ranking; and outputting said optimal preference policy to a participating entity in an auction, said outputting performed prior to conducting said auction.

2. The computer implemented method as recited in claim 1, wherein said selecting characteristics of said auction to be conducted further comprises: receiving a first user input, wherein said first user input comprises information identifying an item to be auctioned; accessing a database; retrieving from said database historical bids data; retrieving from said database auction characteristics data, wherein said auction characteristics comprise information relating to historical auctions of similar items; outputting said bids data; and outputting said auction characteristics data.

3. The computer implemented method as recited in claim 1, wherein said selecting a relevant bidding model for said auction to be conducted further comprises: receiving said auction characteristics data; accessing a database; retrieving from said database a relevant bidding model, wherein said bidding model is selected based on a corresponding relevance of said auction characteristics data; and outputting said relevant bidding model.

4. The computer implemented method as recited in claim 1, wherein said estimating a structure of said auction to be conducted further comprises: receiving said relevant bidding model; receiving bids data; transforming said bids data to a sample of inverted bids, wherein said bids data are transformed by inverting said bidding model; estimating an estimated latent structure of said market, wherein said sample of inverted bids receives application of statistical density estimation techniques to obtain said estimated structure; and outputting said estimated structure.

5. The computer implemented method as recited in claim 1, wherein said bidding model has embedded an unknown structure, and wherein said predicting a bidding behavior for said auction to be conducted further comprises: receiving said estimated structure; receiving said relevant bidding model; substituting said estimated structure for said unknown structure; and outputting a prediction of bidding behavior.

6. The computer implemented method as recited in claim 4, wherein said predicting a first outcome of said auction to be conducted further comprises: receiving a first user input, wherein said first user input comprises: an evaluation criterion; a candidate preference policy; and a constraint; receiving said estimated structure; receiving said bidding behavior prediction for said candidate preference policy, wherein said bidding behavior prediction further comprises a prediction under said constraint; obtaining a value of said evaluation criterion, wherein said value is based on said estimated structure, said bidding behavior prediction, said candidate preference policy, and said constraint, said value comprising said first predicted outcome; and outputting said value.

7. The computer implemented method as recited in claim 6, wherein said evaluating said first outcome of said auction to be conducted further comprises: receiving a second user input, wherein said second user input comprises a plurality of candidate preference policies; receiving a predicted outcome for each said candidate preference policy; calculating descriptive statistics for each said candidate preference policy, wherein said descriptive statistics comprise a mean and a variance; ranking each said candidate preference policy with respect to said calculated mean and generating corresponding rankings for said plurality; and outputting said descriptive statistics and said rankings.

8. A computer system comprising: a bus; a memory interconnected with said bus; and a processor interconnected with said bus, wherein said processor executes a method for determining a preference policy for an auction to be conducted, said method comprising: selecting characteristics of said auction to be conducted; selecting a relevant bidding model for said auction to be conducted; estimating a structure of said auction to be conducted, said estimating comprises expressing unobservable variables in terms of observable bids, wherein said unobservable variables are expressed in terms of observable bids by inverting said bid model; predicting a bidding behavior for said auction to be conducted; predicting a first outcome of said auction to be conducted; and evaluating said first outcome of said auction to be conducted, wherein said evaluating comprises: selecting an optimal preference policy from a plurality of candidate preference policies for treating different groups of bidders differently, wherein said optimal preference policy comprises the candidate preference policy within a plurality having the highest ranking; and outputting said optimal preference policy, prior to conducting said auction, to a participant in said auction.

9. The system as recited in claim 8, wherein said selecting characteristics of said auction to be conducted further comprises: receiving a first user input, wherein said first user input comprises information identifying an item to be auctioned; accessing a database; retrieving from said database historical bids data; retrieving from said database auction characteristics data, wherein said auction characteristics comprise information relating to historical auctions of similar items; outputting said bids data; and outputting said auction characteristics data.

10. The system as recited in claim 8, wherein said selecting a relevant bidding model for said auction to be conducted further comprises: receiving said auction characteristics data; accessing a database; retrieving from said database a relevant bidding model, wherein said bidding model is selected based on a corresponding relevance of said auction characteristics data; and outputting said relevant bidding model.

11. The system as recited in claim 8, wherein said estimating a structure of said auction to be conducted further comprises: receiving said relevant bidding model; receiving bids data; transforming said bids data to a sample of inverted bids, wherein said bids data are transformed by inverting said bid model; estimating an estimated latent structure of said market, wherein said sample of inverted bids receives application of statistical density estimation techniques to obtain said estimated structure; and outputting said estimated structure.

12. The system as recited in claim 8, wherein said bidding model has embedded an unknown structure, and wherein said predicting a bidding behavior for said auction to be conducted further comprises: receiving said estimated structure; receiving said relevant bidding model; substituting said estimated structure for said unknown structure; and outputting a prediction of bidding behavior.

13. The system as recited in claim 11, wherein said predicting a first outcome of said auction to be conducted further comprises: receiving a first user input, wherein said first user input comprises: an evaluation criterion; a candidate preference policy; and a constraint; receiving said estimated structure; receiving said bidding behavior prediction for said candidate preference policy, wherein said bidding behavior prediction further comprises a prediction under said constraint; obtaining a value of said evaluation criterion, wherein said value is based on said estimated structure, said bidding behavior prediction, said candidate preference policy, and said constraint, said value comprising said first predicted outcome; and outputting said value.

14. The system as recited in claim 13, wherein said evaluating said first outcome of said auction to be conducted further comprises: receiving a second user input, wherein said second user input comprises a plurality of candidate preference policies; receiving a predicted outcome for each said candidate preference policy; calculating descriptive statistics for each said candidate preference policy, wherein said descriptive statistics comprise a mean and a variance; ranking each said candidate preference policy with respect to said calculated mean and generating corresponding rankings for said plurality; and outputting said descriptive statistics and said rankings.

15. A computer readable medium having encoded therein a computer readable code for causing a computer system to execute a computer implemented method for determining a preference policy for an auction to be conducted, said method comprising: selecting characteristics of said auction to be conducted; selecting a relevant bidding model for said auction to be conducted; estimating a structure of said auction to be conducted, said estimating comprises expressing unobservable variables in terms of observable bids, wherein said unobservable variables are expressed in terms of observable bids by inverting said bid model; predicting a bidding behavior for said auction to be conducted; predicting a first outcome of said auction to be conducted; and evaluating said first outcome of said auction to be conducted, wherein said evaluating comprises: selecting an optimal preference policy from a plurality of candidate preference policies for treating different groups of bidders differently, wherein said optimal preference policy comprises the candidate preference policy within a plurality having the highest ranking; and outputting said optimal preference policy to a participant in said auction, said outputting performed prior to conducting said auction.

16. The computer readable medium as recited in claim 15, wherein said selecting characteristics further comprises: receiving a first user input, wherein said first user input comprises information identifying an item to be auctioned; accessing a database; retrieving from said database historical bids data; retrieving from said database auction characteristics data, wherein said auction characteristics comprise information relating to historical auctions of similar items; outputting said bids data; and outputting said auction characteristics data.

17. The computer readable medium as recited in claim 15, wherein said selecting a relevant bidding model further comprises: receiving said auction characteristics data; accessing a database; retrieving from said database a relevant bidding model, wherein said bidding model is selected based on a corresponding relevance of said auction characteristics data; and outputting said relevant bidding model.

18. The computer readable medium as recited in claim 15, wherein said estimating further comprises: receiving said relevant bidding model; receiving bids data; transforming said bids data to a sample of inverted bids, wherein said bids data are transformed by inverting said bid model; estimating an estimated latent structure of said market, wherein said sample of inverted bids receives application of statistical density estimation techniques to obtain said estimated structure; and outputting said estimated structure.

19. The computer readable medium as recited in claim 15, wherein said bidding model has embedded an unknown structure, and wherein said predicting a bidding behavior further comprises: receiving said estimated structure; receiving said relevant bidding model; substituting said estimated structure for said unknown structure; and outputting a prediction of bidding behavior.

20. The computer readable medium as recited in claim 18, wherein said predicting a first outcome further comprises: receiving a first user input, wherein said first user input comprises: an evaluation criterion; a candidate preference policy; and a constraint; receiving said estimated structure; receiving said bidding behavior prediction for said candidate preference policy, wherein said bidding behavior prediction further comprises a prediction under said constraint; obtaining a value of said evaluation criterion, wherein said value is based on said estimated structure, said bidding behavior prediction, said candidate preference policy, and said constraint, said value comprising said first predicted outcome; and outputting said value.

21. The computer readable medium as recited in claim 20, wherein said evaluating said first outcome further comprises: receiving a second user input, wherein said second user input comprises a plurality of candidate preference policies; receiving a predicted outcome for each said candidate preference policy; calculating descriptive statistics for each said candidate preference policy, wherein said descriptive statistics comprise a mean and a variance; ranking each said candidate preference policy with respect to said calculated mean and generating corresponding rankings for said plurality; and outputting said descriptive statistics and said rankings.

Description Submit all comments and votes

BACKGROUND OF THE INVENTION

1. Field of the Invention

The present invention relates to the field of computer-based auction design and analysis processes. Specifically, the present invention relates to a method and system for setting an optimal price preference policy for an auction.

2. Related Art

Modern electronic forum based auctions, such as World Wide Web and other Internet based auctions have complex rules with varied and observable characteristics and situations, as well as unobservable structural elements. Auction participants, either sellers or buyers, must make a number of decisions relating to the auction.

Sellers, for example, conducting an auction to sell an item, can improve the auction outcome in their favor by treating bidders with identifiable differences differently. Buyers, correspondingly conducting an auction to buy an item, can also improve the auction outcome in their favor by treating bidders with identifiable differences differently.

Participants entering a market, such as bidders in an auction, differ greatly across a wide spectrum of dimensions. Other market participants, large scale purchasers such as governments or large scale sellers such as major corporations, for example, deal with different bidders entering a market in a variety of differing ways. Illustratively, the United States government offers a 6% price preference for domestically produced U.S. products under legislation mandating what is commonly known as a "Buy-American" policy. Governments of the various states and of other nations have similar policies.

The operation of these price preference policies may be illustrated by the following example. The U.S. Department of Defense offers a 50% price preference to U.S. domestic firms bidding to supply Defense Department purchases. Non-U.S. bidders are at a daunting bidding disadvantage in this situation. Foreign bidders are discriminated in favor of the substantially preferred domestic U.S. firms. If any domestic U.S. supplier's bid is no more than 50% higher than the lowest foreign bid, the domestic bid is accepted. In other words, the preferred domestic U.S. supplier wins in any such auction with the U.S. Department of Defense against a foreign bidder who, without the preference policy in place, would win with a bid of nearly half the sale price.

In business-to-business settings, often less legislatively constrained than governmental market situations, such preferential treatment of some suppliers, with corresponding discrimination against others, is even more prevalent. Similarly, in many business-to-consumer situations, a seller may wish to treat some segment of customers, sharing some particular trait, differently from others. For example, certain customers may be treated preferentially by businesses and other customers discriminatorily.

Illustratively, "loyal" customers, e.g., customers with frequent or repeated significant orders, bidders with better bidder ranking criteria, e.g., higher eBay.RTM. ratings, and customers with identifiably more elastic demands, etc., may be treated preferentially by awarding them a price discount. Similarly, mortgage customers or other borrowers with excellent credit ratings may be awarded a lower interest rate. Conversely, new, e.g., unknown customers, inflexibly rigid customers with stringent accommodation demands, or borrowers with lower credit ratings may represent to a business a higher cost or degree of risk in dealing with them. Such riskier or costlier customers may be discriminated against with higher interest rates, requiring premium prices, or in other handicapping ways.

Setting price preference policies in markets, particularly in auctions, can improve the market outcome in favor of the policy setting market participant, and is thus an important, perhaps crucial business consideration. Currently, these decisions are made by auction participants on an ad hoc basis, sometimes with the assistance of consultants operating themselves on a more or less ad hoc basis. A high degree of uncertainty intrinsic in auction price preference policy related decision making often precludes optimal outcomes, because the soundness of a particular decision in a particular situation cannot be ascertained prior to observation of the outcome (e.g., after the transaction has taken place). Inexperienced auction participants often make unsophisticated sub-optimal decisions regarding the setting of a price preference policy. Experience and a host of other human elements may thus effect the soundness of decision making in a given auction price preference situation. Nevertheless, no conventional systematic auction price preference analytical decision making mechanism is available.

Currently, the decisions on the parameters of preference policy are left entirely to the person conducting the auction. There is little systemic data analysis to guide these decisions. Given the multiplicity of items bought/sold through auctions, it is typically too costly to hire expert analysis to configure the price-preference policies for each case. Typically, a given policy, say 10% preference for preferred suppliers, is applied to a large class of procurement situations. Yet bidders' cost distributions vary considerably across procurement items and across time. A fixed preference policy is rarely optimal for every case to which it is applied.

As is known, the outcome of an auction (e.g., who gets what, who pays how much) is determined by bidding behavior of bidders. Bidding behavior depends on, among other factors, the auction rules in that different auction rules induce different behavior on the part of the bidders. A bidder's behavior under a given collection of auction rules in turn is determined by the bidder's private information. The structure of the private information held by the bidders is thus a key factor in evaluating alternative auction rules. This fundamental element of the auction environment is not directly observable and has to be estimated from observable and available data.

There exists a need for an automated estimation and optimization solution for configuring the parameters of preference policies to be implemented in auctions. What is needed is a method and/or system that configures the optimal preference policies that can be combined with any auction format a market decision maker may wish to conduct. What is also needed is a method and/or system that applies to any auction participants, either buyers conducting an auction to procure an item, or a seller, conducting to sell an item, which estimate's bidders private information and correspondingly identifies exploitable asymmetries. Further, what is needed is a method and/or system that achieves the foregoing to implement a preferential treatment policy.

SUMMARY OF THE INVENTION

An embodiment of the present invention provides a method and system that configures the optimal preference policies that can be implemented in any market, particularly an auction, that can be combined with any auction format a market decision maker may wish to conduct. An embodiment of the present invention provides a method and system that applies to any auction participants, either buyers conducting an auction to procure an item, or a seller, conducting an auction to sell an item, which estimates bidders' private information and correspondingly identifies exploitable asymmetries. Further, an embodiment of the present invention provides a method and system that achieves the foregoing to implement a preferential treatment policy.

In one embodiment, the present invention provides a method and system that estimates bidder's private information, and correspondingly identifies exploitable asymmetries between market participants by which preferential treatment policies may be implemented. In one embodiment, structural analysis of bid data from prior auctions is used to identify and estimate the distributions of bidders' private signals, conditional on observable bidder characteristics.

In one embodiment, the estimated distributions of bidders' signals, identified by structural analysis, are examined to identify significant asymmetries across the population of bidders that can be used to personalize, e.g., to particularize, design parameters for the auction to be conducted. In one embodiment, these parameters include, but are not limited to, reserve prices, entry fees, winner determination rules, and payment rules of the auction to be conducted.

In one embodiment, a computer system executes, under the control of software and firmware directing the operation of its processor and other components, a process that estimates bidder's private information, and correspondingly identifies exploitable asymmetries between market participants by which preferential treatment policies may be implemented.

In one embodiment, a computer readable medium causes a computer system to execute the steps in a method for implementing a process that estimates bidder's private information, and correspondingly identifies exploitable asymmetries between market participants by which preferential treatment policies may be implemented.

These and other objects and advantages of the present invention will become obvious to those of ordinary skill in the at after reading the following detailed description of the preferred embodiments which are illustrated in the drawing figures.

BRIEF DESCRIPTION OF THE DRAWINGS

The accompanying drawings, which are incorporated in and form a part of this specification, illustrate embodiments of the invention and, together with the description, serve to explain the principles of the invention.

FIG. 1 schematically shows an automated decision support system for designing auctions in accordance with an embodiment of the invention.

FIG. 2 shows in block diagram form the structure of the structure extractor of the decision support system of FIG. 1.

FIG. 3 shows in block diagram form the structure of the behavior predictor of the decision support system of FIG. 1.

FIG. 4 is a flow chart of steps in a process for generating auction characteristics data, in accordance with one embodiment of the present invention.

FIG. 5 is a flow chart of steps in a process for generating a relevant bidding model, in accordance with an embodiment of the present invention.

FIG. 6 is a flow chart of steps in a process for generating an estimated market structure, in accordance with an embodiment of the present invention.

FIG. 7 is a flow chart of steps in a process for predicting bidder behavior, in accordance with an embodiment of the present invention.

FIG. 8 is a flow chart of steps in a process for determining an optimal preference policy, in accordance with an embodiment of the present invention.

FIG. 9 is a flow chart of steps in a process for reporting preference policy ranking, in accordance with an embodiment of the present invention.

FIG. 10 is a block diagram depicting a computer system and computer readable media for implementing processes of market preference policy determination, in accordance with an embodiment of the present invention.

FIG. 11 (sheets 1-4) depicts contents of a database of market data, in accordance with an embodiment of the present invention.

FIG. 12 is a flow chart of steps in a process for determining an optimal reserve price for an auction, in accordance with an embodiment of the present invention.

DETAILED DESCRIPTION OF THE INVENTION

In the following detailed description of the present invention, numerous specific details are set forth in order to provide a thorough understanding of the present invention. However, it will be recognized by one skilled in the art that the present invention may be practiced without these specific details or with equivalents thereof. In other instances, well known methods, procedures, components, and circuits have not been described in detail as not to unnecessarily obscure aspects of the present invention.

Notation and Nomenclature

Some portions of the detailed descriptions, which follow, are presented in terms of procedures, steps, logic blocks, processing, and other symbolic representations of operations on data bits that can be performed by computer systems. These descriptions and representations are used by those skilled in the data processing arts to most effectively convey the substance of their work to others skilled in the art. A procedure, computer executed step, logic block, process, etc., is here, and generally, conceived to be a self-consistent sequence of steps or instructions leading to a desired result. The steps are those requiring physical manipulations of physical quantities. Usually, though not necessarily, these quantities take the form of electrical, electronic, magnetic, optical, and/or electro-optical signals capable of being stored, transferred, combined, compared, and otherwise manipulated in a computer system. It has proven convenient at times, principally for reasons of common usage, to refer to these signals as bits, values, elements, symbols, characters, terms, numbers, or the like.

It should be borne in mind, however, that all of these and similar terms are to be associated with the appropriate physical quantities and are merely convenient labels applied to these quantities. Unless specifically stated otherwise as apparent from the following discussions, it is appreciated that throughout the present invention, discussions utilizing terms such as "accessing" or "calculating" or "constraining" or "estimating" or "evaluating" or "expressing" or "inputting" or "outputting" or "predicting" or "ranking" or "receiving" or "retrieving" or "selecting" or "substituting" or "transforming" or "promulgating" or the like, refer to the action and processes of a computer system, or similar electronic computing device, that manipulates and transforms data represented as physical (electronic) quantities within the communications and computer systems' registers and memories into other data similarly represented as physical quantities within the computer system memories or registers or other such information storage, transmission, or display devices.

Certain portions of the detailed descriptions of the invention, which follow, are presented in terms of processes (e.g., process 40, FIG. 4). These processes are, in one embodiment, carried out by processors and electrical and electronic components under the control of computer readable and computer executable instructions. The computer readable and computer executable instructions reside, for example, in registers and other features of processors, memories, and data storage features of computers executing programs and processes. However, the computer readable and computer executable instructions may reside in any type of computer readable medium. Although specific steps are disclosed in figures herein describing the operations of processes (e.g., FIG. 4; describing process 40), such steps are exemplary. That is, the present invention is well suited to performing various other steps or variations of the steps recited in the flowcharts of the figures herein. Further, it is appreciated that the steps of these processes may be performed by software or hardware or any combination of software and hardware.

The present invention is discussed primarily in the context of a method for determining an optimal preference policy for a market structure, such as an auction, with respect to a multiplicity of possible market participants, such as auction participants like sellers and bidders, or other end users specify. In the following discussion, a market analysis system, especially applicable to auctions will first be discussed in a general context, after which, specific explanations will be made to applying the system to selecting optimal preference policies.

Exemplary System

Exemplary Automated Decision Support System in General

FIG. 1 shows an automated decision support system 10 for designing auctions in accordance with one embodiment of the present invention. In one embodiment, the automated decision support system 10 is a software system implemented in a computer system. Alternatively, the automated decision support system 10 can be implemented by hardware or firmware.

The computer system that embodies the automated decision support system 10 can be any kind of computer system. For example, the computer system can be a main-frame computer system, a super computer system, a workstation computer system, a server computer system, or a personal computer system.

The automatic decision support system 10 is used to provide decision support for auction design. This means that the automatic decision support system 10 aids auction sellers, buyers, bidders, or auction houses in making auction-related decisions. As described above, there are typically a number decisions to be made regarding an auction. For example, in an auction run by a seller, a bidder has to decide on (1) how to bid and (2) whether or not to bid in a specific auction conditional on information the bidder has. In addition, the bidder needs to decide whether or not and how to gather information on auctions, objects, rivals. The auction house for the auction needs to decide fees for buyers and/or sellers. In addition, the auction house needs to decide the menu of auction mechanism to offer.

Similarly, in an auction run by a seller, the seller also has to decide what the reserve price of the auctioned item should be, what is the best auction format, what entry fees should be charged for participating in the auction, what timing and duration of the auction should be, the quantity of the item to be auctioned, what participation rules should govern the auction, and what information rules should be imposed to the auction, etc. As is known, these decisions affect the revenue or profit generated from the auction.

In order to achieve the maximum revenue or profit, these decisions must be optimized. In accordance with an embodiment of the present invention, this optimization is done by automatic decision support system 10. Automatic decision support system 10 provides optimal configuration of auction design parameters and comparative evaluation of any pair of design choices. In other words, automatic decision support system 10 provides automated auction analysis optimization.

In accordance with an embodiment of the present invention, automatic decision support system 10 processes available data using structural econometric techniques to identify latent distribution of random or unknown elements of a market structure or market environment of a particular auction. In addition, automatic decision support system 10 provides the optimal values of any subset of the decision variables or candidates based on an evaluation criterion specified by a user of system 10 conditional on levels of the remaining decisions.

The available data to the automatic decision support system 10 include data supplied by the user of the system 10. The data include description of the item to be auctioned, auction decision candidates, constraints, and auction evaluation criterion. These are user inputs to the automatic decision support system 10. The available data also include historical auction data and bidding model data. The historical auction data and the bidding model data are stored in the automatic decision support system 10.

The automatic decision support system 10 receives the user inputs of the description of the item to be auctioned, the auction decision candidates, the constraints, and the auction evaluation criterion. The automatic decision support system 10 then selects the best auction decision candidates (e.g., the best auction format is English, the reserve price is $100, the entry fee is $5, and the duration is five days) among the inputted auction decision candidates based on the auction evaluation criterion and the estimated market structure of the auction.

The market structure affects bidding behavior of bidders during the auction. As is known, bidding behavior determines the outcome of an auction. The outcome of an auction means who gets what and who pays how much, etc. Different auction rules induce different bidding behavior on the part of bidders. A bidder's behavior under a given set of auction rules in turn is determined by his private information. The structure of private information held by bidders is thus a key factor in evaluating alternative auction procedures (e.g., auction format, reserve prices, entry fees, timing and duration of the auction, quantity, participation rules, and information rules, etc.) from the point of view of a seller (or buyer) trying to sell (or procure) an item by auctioning. This fundamental element of an auction environment is not directly observable and has to be estimated from observable and available data. The auction procedures can also be referred to as auction mechanisms. They include the characteristics of the auction.

In accordance with an embodiment of the present invention, automatic decision support system 10 estimates the unknown or unobservable elements of the market structure of the auction by extracting the joint distribution of private information of the bidders (e.g., the probability distribution of bidders' willingness to pay, the probability distribution of the number of potential bidders) from bid data extracted from the historical auction data of similar auctions. In particular, automatic decision support system 10 estimates the unknown elements of the market structure by (1) expressing unobservable variables in the bidding model in terms of the observable bid data, and (2) applying known statistical density estimation techniques to the expression so as to obtain an estimation of the unknown elements. In doing so, automatic decision support system 10 enables the user (either a seller or a buyer) of system 10 to factor the distribution of bidders' private information into his or her decisions regarding the appropriate auction procedure to conduct the auction.

With the estimated unknown elements of the market structure and other user inputs (e.g., the auction design candidates, evaluation criterion), the automatic decision support system 10 provides optimized auction design candidates based on the evaluation criterion provided such that maximized expected revenue or profit from the auction can be achieved. This means that the automatic decision support system 10 can be used to configure optimized auction parameters for a multiplicity of performance criteria. The structure and operation of the automatic decision support system 10 will be described in more detail below, also in conjunction with FIGS. 1 through 10.

As can be seen from FIG. 1, automatic decision support system 10 includes a historical auction data repository 11, a bidding model repository 12, a structure extractor 13, a behavior predictor 14, and an optimizer 15. The historical auction data repository 11 stores historical auction data for previous auctions. The historical auction data specify auction characteristics and/or mechanisms of previous auctions. This means that the historical auction data includes the bid data and the auction characteristics data of each of the stored previous auctions. The auction characteristics data specify the auction procedure of the auction. Thus, the auction characteristics data of an auction describe the reserve price of the auctioned item, the auction format, the number of bidders, etc. of the particular auction. The bid data of an auction describe the bidding behaviors of bidders in the auction. The bid data is a record that typically contains the auction identifier, number of bidders N, number of bids, transaction price, winner, reserve price, auction format, item characteristics, and bidder characteristics. Both the bid data and the auction characteristics data are extracted from the auction data of the previous or historical auctions for various items. The historical auction data repository 11 can be implemented using any known database technology.

The bidding model repository 12 stores various bidding models. A bidding model specifies a bidding behavior pattern. It is a function of auction characteristics or procedure of the corresponding auction. It is also a function of the market structure of the auction. For example, a Dutch auction bidding model specifies bidding behavior in a Dutch format auction. An English auction bidding model specifies bidding behavior in an English format auction. A first-price-sealed-bid auction bidding model specifies bidding behavior in a first-price-sealed-bid auction. The bidding model repository 12 can be implemented using any known database technology. Several exampl