KR20150057983A - Search engine for identifying business travel proposals - Google Patents

Abstract

The present invention provides systems, methods, and computer program products for ranking and selecting travel-related search results for display to a traveler. The system may receive data identifying a preferred travel proposal from each of a plurality of travelers, and analyze the data to determine a value utility function. The value utility function may then be used to assign a relative utility to travel proposals identified in response to a search query from a traveler. Additional data relating to a pricing sensitivity of a payer may be received from an entity paying for travel, such as a corporation or other business entity. The system may analyze the pricing sensitivity data to define a price utility function that quantifies the price of each travel proposal. A composite utility function that is a weighted sum of the value and price utility functions may be used to select search results for display to the traveler.

Description

SEARCH ENGINE FOR IDENTIFYING BUSINESS TRAVEL PROPOSALS

Field of the Invention The present invention relates generally to computer and computer software, and more particularly to a method, system, and computer program product for providing search results in response to search queries for travel-related products.

Travelers who participate in business-related travel for businesses or other business entities can search and book travel goods through a computer reservation system. In order to find a trip satisfying a traveler's trip requirement, the traveler typically provides search terms in the form of a search query to the reservation system. In response to receiving the search query, the booking system may search the one or more databases to identify the travel goods matching the parameters in the search query. The matching travel product may be displayed as part of one or more travel proposals and booked or purchased by the traveler.

Depending on the search item provided by the traveler, the search engine of the reservation system may return multiple travel offers. In general, search results are provided as a list of travel suggestions, only a portion of which is displayed at one time. If the traveler does not find a satisfactory travel offer in the initially displayed search results, the traveler may scroll down the list or request that another portion of the list be displayed. Thus, travelers are presented with the task of sorting a number of travel offers to find travel offers that meet the traveler's needs.

Travelers overwhelmed by the number of travel offers presented may have difficulty choosing a travel offer, or may choose to opt for a non-optimal travel offer near the top of the list, giving up searching for the best travel offer in the list. Search results that present too many options to travelers can confuse travelers and increase the time it takes to complete the bookkeeping process. Many irrelevant options, such as travel offers, including flights, that do not fit the schedule or are too verbose, may be presented to business travelers. A business entity that pays for travel may prefer that an employee follow a corporate travel policy and / or select a travel option with a lower negotiated amount. At the same time, the company may want employees to be satisfied, taking into account employee preferences.

Accordingly, there is a need for an improved system, method, and computer program product that improves the ability of a traveler to travel to search for and book a travel offer that identifies a travel offer that meets both the needs of the business entity and traveler's needs.

In one embodiment of the invention, a method is provided for featuring a plurality of travel offers that satisfy a travel request. The method includes receiving first data identifying a preferred travel offer for each of a plurality of travelers. Each preferred travel offer is selected from a plurality of travel offers, and the first data is analyzed to define a first utility function that quantifies the value of each travel offer. The method includes receiving second data on price sensitivity from a payer and defining a second utility function for analyzing the second data to quantify the price of each travel offer . A third utility function having an output that quantifies the combined utility for each travel offer may be defined as a weighted sum of the first and second utility functions.

In another embodiment of the present invention, an apparatus is provided that features a plurality of travel offers that satisfy a travel request. The apparatus includes a processor and a memory operably coupled to the processor. Wherein the memory comprises program instructions and the program instructions, when executed by the processor, cause the device to perform the operations of receiving first data identifying a preferred travel offer for each of the plurality of travelers, Each preferred travel offer is selected from the plurality of travel offers. The apparatus defines a first utility function for analyzing the first data to quantify the value of each travel offer. The device further receives second data on price sensitivity from the payer and analyzes the second data to define a second utility function to quantify the price of each travel offer. The third utility function is defined as a weighted sum of the first and second utility functions by the device to have an output that quantifies the hybrid utility for each travel offer.

In yet another embodiment of the present invention, a computer program product is provided. The computer program product includes a non-transitory computer readable storage medium for storing program instructions. Wherein the program instructions are configured to, when executed by the processor, cause the processor to perform the operation of receiving first data identifying a preferred travel offer for each of the plurality of travelers, each preferred travel offer comprising a plurality ≪ / RTI > The program command causes the processor to analyze the first data to define a first utility function that quantifies the value of each travel offer. The program instructions causing the processor to perform operations to receive second data on price sensitivity from the payer and to define a second utility function to analyze the second data to quantify the price of each travel offer Lt; / RTI > The program instructions may cause the processor to perform an operation to define a third utility function having an output that quantifies the complex utility for each travel suggestion function as a weighted sum of the first and second utility functions .

BRIEF DESCRIPTION OF THE DRAWINGS The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate various embodiments of the invention and, together with the general description of the invention given above and the detailed description of the embodiments given below, It plays a role.
1 is a schematic diagram of an exemplary operating environment including a computer reservation system for communicating with a plurality of computer systems via a network.
Figure 2 is a schematic diagram of the exemplary computer system of Figure 1;
Figure 3 is a schematic diagram of a graph showing a plurality of iso-utility curves.
Figure 4 is a schematic diagram of a process in which the reservation system of Figure 1 can select a travel offer to display to a traveler.
Figure 5 is a schematic diagram of a graph including a utility function curve defining a price utility function that can be used by the process of Figure 4;
6 is a flow chart illustrating a process by which a reservation system can determine a parameter weight for a value utility function that can be used by the process of FIG.
7 is a flow chart illustrating a process by which a reservation system can determine an optimal number of typical traveler profiles.

Embodiments of the present invention are directed to a system and method for providing a user with travel-related search results. In particular, the reservation system may provide search results for trips where the traveler is not the paying side or the "payer ". One example where travelers and payers may not be the same entity is a business trip. In a business trip, businesses or other business entities typically pay travel expenses for traveling employees. In order to identify travel offers that satisfy the search query, the reservation system may search for travel-related goods and services from a number of sources. These sources may include a Global Distribution System (GDS), a travel goods provider such as an airline and hotel, a third party travel product provider, an online travel agent, or any other suitable database of travel-related goods.

In order to select and display the most relevant search results to the traveler, the reservation system can extract the value preferences from the traveler and extract the cost preferences from the payer. These preferences can be used to define a composite utility function that determines the relative utility of the travel offer returned by the search query. The composite utility function can be used to rank the travel offers returned by the search query to display the most relevant search results to the traveler. The reservation system can take into account both the payer and the traveler's preferences in determining the search results to display to the traveler.

To this end, each search result includes (1) a price utility function that quantifies the price of a travel product based on a payer's preference; And (2) a value utility function that quantifies the value of the travel goods for the traveler. The composite utility value for each search result can be determined based on the weighted sum of the quantified price and value. The search results may be selected to display to the traveler based on this composite utility.

1, an operating environment 10 in accordance with one embodiment of the present invention includes a user system 12, one or more merchandise provider systems 14, and a reservation system 16, . ≪ / RTI > The network 18 may include one or more private and / or public networks (e.g., the Internet) that are capable of exchanging data between systems. In one embodiment of the present invention, the reservation system 16 may be part of a Global Distribution System (GDS), a travel management system, or any other suitable computer system used to book and / or book a travel product.

User system 12 may be a desktop computer, a laptop computer, a tablet computer, a smart phone, or any other computing device that provides the user with access to the system 12, 14, 16 connected to the network 18. Merchandise provider system 14 and / or reservation system 16 may each host one or more databases that store information about the travel goods. These databases may include data relating to travel goods provided by an online travel agent, a meta-search engine, an airline, a hotel, a GDS, or any other provider or seller of travel goods. The merchandise provider system 14 may be configured to provide the booking system 16 with data regarding the travel goods. This data may be provided from reservation system 16 in response to a database query. In this way, the traveler can retrieve and book the travel goods marketed by the goods provider via the reservation system 16. The reservation system 16 may further include a database of travel goods. The reservation system 16 allows a user to search, book, and / or book a trip through an appropriate software application, such as a web browser, running on the user system 12. The reservation system 16 is maintained by a business entity, a travel management company, or an operator of the GDS, and can manage employee travel for business entities.

2, user system 12, merchandise provider system 14, and reservation system 16 of operating environment 10 may be implemented in one or more computer devices or systems, such as exemplary computer system 22, . The computer system 22 may include a processor 24, a memory 26, a mass storage medium 28, an input / output (I / O) interface 30, and a user interface 32.

The processor 24 may be a microprocessor, a micro-controller, a digital signal processor, a microcomputer, a central processing unit, an electric field programmable gate array, a programmable logic device, a state machine, a logic circuit, Or any other device that manipulates the signal (analog or digital) based on an operating command stored in the memory (not shown). The memory 26 may be a read only memory (ROM), a random access memory (RAM), a volatile memory, a non-volatile memory, a static random access memory (SRAM), a dynamic random access memory But may include a single memory device or a plurality of memory devices, which may include, but are not limited to, any other device capable of storing information. The mass storage media 28 may include a hard drive, an optical drive, a tape drive, a data storage device such as a non-volatile solid state device, or any other device capable of storing information. The database 34 may reside on mass storage media 28 and may be used to collect and organize data used by the various systems, applications, modules, and tools described herein.

The processor 24 may operate under the control of an operating system 36 resident in the memory 26. The operating system 36 may include computer program code or instructions embodied in one or more computer software applications, such as applications 38 residing in memory 26, such as computer resources Can be managed. In an alternative embodiment, the processor 24 may execute the application 38 directly, in which case the operating system 36 may be omitted. One or more data structures 40 may reside in memory 26 and may be used by processor 24, operating system 36, and / or application 38 to store or manipulate data.

The I / O interface 30 may provide a machine interface operatively connecting the processor 24 to other devices and systems, such as the network 18. The application 38 may communicate through the I / O interface 30 and may provide various features, functions, and / or modules, including embodiments of the invention, in cooperation with the network 18. The application 38 may further include program code that is executed by one or more external resources (not shown) or may be dependent upon functions and / or signals provided by other systems or network components external to the computer system 22 . In fact, hardware and software configurations can be provided almost endlessly, so those of ordinary skill in the art will appreciate that embodiments of the present invention may be located external to computer system 22 or may be distributed to multiple computers or other external resources Or may be provided by an application provided by a computing resource (hardware and software) served through a network 18, such as a cloud computing service.

The user interface 32 is operatively connected to the processor 24 of the computer system 22 in a known manner to enable the user to interact directly with the computer system 22. [ The user interface 32 may include video and / or alphanumeric displays, touch screens, speakers, and any other suitable audible, visual, and / or tactile indicators capable of providing information to the user . The user interface 32 may include input devices such as alphanumeric keyboards, pointing devices, keypads, push buttons, control knobs, microphones, etc., which may receive commands or inputs from a user and communicate input to the processor 24 and / And a control unit.

Referring now to FIG. 3, an exemplary graph 50 illustrates the two-dimensional representation of the utility for a plurality of travel offers represented by markers 52-54. The graph 50 includes a vertical axis 58 representing the price utility of the travel offer and a horizontal axis 60 representing the value utility of the travel offer. Each of the price and value utilities may have a minimum value of 0% or 0, and a maximum value of 100% or 1, where 0 represents the lowest utility (e.g., the highest possible price or lowest value) (For example, the lowest possible price or the highest value). The upper and lower bounds of the input to the utility function may be determined based on a travel offer that satisfies a particular type of search query, such as a travel query, between a certain source-destination city pair. That is, the upper and lower limits may be set to correspond to the highest and lowest value and price among the set of travel suggestions that satisfy a particular question. The upper and lower limits may be determined based on a set of travel suggestions that includes a plurality of travel searches for all travel offers in the database, or any other suitable travel offer set. The upper and lower limits may be determined based on the selected value. For example, the highest value may represent a hypothetical travel offer with parameters selected to match all the traveler's preferences, even though the travel offer matching all the preferences may not be actually available. As an example, the marker 52 may represent the hypothetically worst travel offer with the highest price and lowest value possible. Marker 54 may represent the hypothetically best travel offer with the highest possible value and the lowest price possible. Thus, the price and value utility of the marker 52 may be (0, 0), and the price and value utility of the marker 54 may be (1, 1).

The location of the markers 52-54 in this graph 50 may be determined by a price utility function that quantifies the price of the corresponding travel offer and a value utility function that quantifies the value of the corresponding travel offer. The price and value utility function may map each travel offer to a location on the graph (50). This mapping may be based on the price determined from the data derived from the payer, and on the value of the travel offer determined from the data derived from the one or more travelers. The combined utility for travel offers can be determined based on the weighted sum of price and value utilities. The composite utility can be used to determine a travel offer to display to a traveler from a plurality of search results.

The graph 50 further includes a plurality of isoform curves 62-68 representing points having an equivalent complexity in the graph 50. [ That is, the weighted sum of price utility and value utility can be constant for all points on the iso-utility curve (62-68). As can be seen from the iso-efficiency curve (62-68), when the price utility increases (i.e., when the price of the travel offer drops), the value utility decreases and generally reduces the value of the lower- . ≪ / RTI > In the graph 50 shown, the iso-efficiency curve 62-68 may exhibit a compound effect that is generally lower for points on the graph 50 and generally increasing for the points on the right. Thus, the travel offer indicated by the marker 53 has a higher combined utility than the travel offer indicated by the marker 52, and may have a lower composite utility than the travel offer indicated by the marker 54. [

Anyway, a travel offer with the highest combined utility is not always a travel offer with the lowest price or highest value to the traveler. Travel offers with the highest combined utility can vary depending on the weight of the price and value utility function. In general, neither the worst travel offer nor the best travel offer is actually available. Rather, the reservation system 16 can hypothetically attempt to take as close as possible to the best travel offer by finding the best available compromise between price and value.

The value utility may be determined based on a weighted sum of a plurality of traveler preferences:

(수식 1)

(Equation 1)

Where v is the value utility, w _c is the weight applied to the corresponding parameter of the travel offer, u _c is the utility function for the parameter, and v _c is the value of the parameter. The utility function u _c may have a binary output depending on the presence or absence of a parameter (e.g., the presence of an airline rewards program), or may be based on the value of the parameter (e.g., flight time) And may have scaled outputs between 0 and 1. In an exemplary embodiment, the weights may be selected such that the sum of the weights is 1, as in Equation 2:

(수식 2)

(Equation 2)

Using the weight (w _c ) in Equation 2, a travel offer with a parameter set that fully satisfies all traveler preferences can have a value utility of one. A travel proposal with a parameter set that is partially unsatisfactory to any of the traveler's preferences can have a value utility of O. Examples of parameters that can be used to determine the value utility of a travel offer include flight length, flight arrival time, rating, distance between hotel and meeting place, presence of an airline compensation program, and presence of a hotel compensation program. But are not limited to these. Parameters can be used as constraints to remove travel suggestions from considerations. For example, if the time between the departure time of the last scheduled meeting in a business trip and the flight offer is less than a predetermined period, the travel offer may be excluded from the search results. The minimum allowed time may be a fixed time period, or may be determined based on trip parameters such as the last scheduled meeting location and the distance between the airport or the estimated travel time.

Price utility can be determined using a price utility function (u _p (P)) that can return a utility between 0 and 1 depending on the price (P). The price utility function u _p (P) may be a function expressing the manner in which the payer evaluated the price of the travel offer. The combined utility (U (P, V)) of a travel offer can be determined based on the weighted sum of value utility and price utility as follows:

(수식 3)

(Equation 3)

Equation 3 substitutes V _x into Equation 2 and rewrites it as follows:

(수식 4)

(Equation 4)

Where P _x represents the price of the travel offer (T _x ), V _x represents the value utility of the travel offer (T _x ), and w _p is the price weight or "balance factor".

The value utility can be limited to reflect the traveler's preference regardless of price. The value utility can quantify the extent to which a traveler value a travel offer regardless of price so that the output of the value utility is a function of the value component of the travel offer (v _c ) only.

The price weight (w _p ) may have a numerical value between 0 and 1, which reflects the relative importance of the verses value determined by the payer. The plurality of travel offers, for example, the travel offers returned in response to the search query, may be characterized by evaluating U (P _x , V _x ) for each travel offer (T _x ). The travel offer (T ₁ ) may be preferred to the travel offer (T ₂ ) if U (P ₁ , V ₁ )> U (P ₂ , V ₂ ) The composite utility (U (P _x , V _x )) can be used to determine the travel offer to display to the traveler.

The price utility function (u _p ) can be configured to reflect the payer's sensitivity to price. To this end, the payer may limit the price preference between trip suggestions or "intra-price" preferences that are different than the price range including the estimated travel offer (T _x ). A simple price utility function (u _p ) can have a linear relationship with price. For example, the price of $ 0 can have a utility of 1, and the highest price that a payer is willing to pay can have a utility of zero, in which case the utility function is limited to a line connecting these two points . More complex price utility (u _p ), which has a non-linear relationship between price and utility, may be limited. The payer can define how much weight to place on the verses value. This price weight can provide a balancing factor between the interest of the business entity and the interest that meets the traveler's travel preference when regulating the price.

Referring now to FIG. 4, a process 80 for selecting a travel offer to display to a traveler is presented. The process 80 can be performed by the reservation system 16 and includes a group 82 of one or more processes for deriving information from the payer and a group 84 of one or more processes for deriving information from the one or more travelers, . ≪ / RTI > The derived data may be collected and used to define a utility function used to select a travel offer to display to the traveler.

At block 86, the payer may provide data to the reservation system 16 that identifies the criteria used to define the utility function U (P _x , V _x ). This criterion may include a value parameter and a constraint parameter. The value parameter may include a parameter that can generate an output range when plugged into each utility function (u _c , u _p ). The value parameter may be specified by a specific source-destination pair of cities, e.g., flight length or arrival time, or may be non-specified. Examples of non-specific city pair value parameters may include, but are not limited to, the price of the travel offer, the hotel rating, the distance between the hotel and the scheduled activity location, and the presence or absence of the airline and / Do not. The constraint parameter may be related to the criteria required for a travel offer - that is, the criteria for a travel offer that is required to satisfy the trip requirement. Examples of constraint parameters may include, but are not limited to, airline arrangement, and the time between the last scheduled activity of the trip and the scheduled departure time of the return flight.

For each selected criterion, one or more boundaries may be defined for the parameters of the criterion. These boundaries may include upper and lower bounds. In one embodiment of the invention, a boundary representing the best-case parameter value (e.g., the shortest available flight time) may be assigned a utility of one, and the worst-case parameter (e.g., The longest flight time) may be assigned a utility of one. For a binary criterion (e.g., a criterion containing a parameter having a yes / no value), the desired parameter value may produce a utility of one, and the undesired parameter value may produce utility of O. For parameters with successive values, the utility function u _c (v _c ) may have a linear output between the upper and lower bounds.

To this end, the travel basis utility function u _c (v _c ) has an output of zero for a corresponding reference value v _c at a minimum level (e.g., a lower limit) and at a maximum level (e.g., Upper limit) and a corresponding reference value (v _c ). Equation 5 provides an exemplary parameter utility function for a criterion (v _c ) that a higher reference parameter value correspondingly provides a higher value to the traveler:

(수식 5)

(Equation 5)

Equation 6 below can provide an exemplary parameter utility function for a criterion (v _c ) where a lower reference parameter value provides the traveler with a corresponding higher value such as flight time:

(수식 6)

(Equation 6)

Similarly, the linear price value function may be provided as: < RTI ID = 0.0 >

(수식 7)

(Equation 7)

Other pricing functions with a slope that varies depending on the price and / or price mapping function defined by the payer may also be used.

At block 88, the reservation system 16 may extract the payer's internal price preference based on the payer's response to one or more questions. One type of question can present two travel offers (T ₁ , T ₂ ) to the payer. The travel offer T ₁ may be characterized by a price P ₁ and a value utility V ₁ and the travel offer T ₂ may be characterized by a price P ₂ and a value utility V ₂ have. In one embodiment of the invention, each value utility (V ₁ , V ₂ ) may be associated with a corresponding set of parameters defining the selected criteria of each travel offer (T ₁ , T ₂ ). The selected criteria can be selected so that the travel offer T ₁ has a lower value utility V ₁ and correspondingly lower price P ₁ than the travel offer T ₂ . The information provided to the payer may omit the price or value of one of the travel offers (T ₁ , T ₂ ). For example, the payer for _{T 1 = (P 1, V} 1), T 2 = (P 2, X 2), or _{T 1 = (X 1, V} 1), T 2 = (P 2, V 2 ) May be provided, where X ₁ represents the empty price field and X ₂ represents the bin value field in the data received by the payer.

If the price (V ₂ ) is omitted, the payer requires that the value of V ₂ be limited so that the composite utility (U (P ₁ , V ₁ )) equals the composite utility (U (P ₂ , V ₂ ) Can receive. That is, the reservation system 16 may require the payer to limit the travel option (T ₂ ) to the minimum value (V ₂ ) that may be needed to accommodate the price increase from P ₁ to P ₂ by the payer. Value (P ₁₎ in this case is omitted, the payment shall be equal to the composite utility _{(U (P 1, V 1} )) The compound efficacy _{(U (P 2, V 2} )) required to define the value (P ₁₎ Can receive. That is, the reservation system 16 may require the payer to limit the minimum price difference (P ₂ - P ₁ ) that may be needed to accommodate the value decrease from V ₂ to V ₁ . An embodiment of the present invention may further include the case where V ₁ or P ₂ is omitted from the data and the payer may define these values such that U (P ₁ , V ₁ ) = U (P ₂ , V ₂ ) . Enter the value provided by the payer into Equation 3 as follows:

_{w p xu p (P 1)} + (1-w p) x V 1 = w p xu p (P 2) + (1-w p) x V 2 ( Formula 8)

As described in more detail below, the reservation system 16 may use the value (X _x ) provided by the payer and the output of the price utility function (u _p ) for one or more input prices (P _x ) To define a price utility function (u _p ) for the payer.

By way of example, the price (P ₀ ) may be set to the highest price at which the payer is willing to accept the travel offer, for example, $ 5000. The price (P ₀ ) may be associated with a sample number of 0 in the array, which may be maintained as part of the database. Since this worst case price has zero utility by definition, this can be seen as:

u _p (P ₀ ) = 0 (Equation 9)

Similarly, if the best-case travel offer (T ₀ ) satisfies all traveler preferences (e.g., the shortest flight available and top-rated hotel), then the known value utility is:

V ₀ = 1 (Equation 10)

The worth paying utility (V ₁ = 0) (for example, usable up to flights and low grade hotel) to have travel suggestions (T ₁₎ travel suggestions (T ₀₎ and the difference in price that may have the same composite utility ( δ ₁ ) can be requested. That is, the payer may be required to determine the price (P ₁ = P ₀ - δ ₁ ), which is as follows:

U (P ₀ , 1) = U (P ₁ , 0) (Equation 11)

In response to the query in this example, the payer may provide data indicating that δ ₁ = $ 500, which may be sufficient to satisfy Equation 11 resulting in a price (P ₁ ) of $ 4500. The sample number may be incremented (e.g., from 0 to ₁ ) and associated with price (P ₁ ) in the array.

The process of extracting the internal price reference is to set the worst case price equal to P ₁ so that the reservation system 16 becomes:

u _p (P ₁ ) = 0 (Expression 12)

, It can be continued to select the best case of travel offer satisfying all traveler preferences (e. G., The shortest flight available and top rated hotel) to be the equation 13.

V ₁ = 1 (Equation 13)

The worth paying utility (V ₂ = 0) (for example, usable up to flights and low grade hotel) to have travel suggestions (T ₂₎ travel suggestions (T ₁₎ and the difference in price that may have the same composite utility ( δ ₂ ) can be requested. That is, the payer may be required to determine the price (P ₂ = P _{1 -} δ ₂ ), which is as follows:

U (P ₁ , 1) = U (P ₂ , O) (Equation 14)

As before, the sample number may be incremented (e.g., from 1 to ₂ ) and associated with price (P ₂ ) in the array. As an example, in response to the above question, the payer may provide data indicating that δ ₂ = $ 700, which may be sufficient to satisfy Equation 14 resulting in a price (P ₂ ) of $ 3800.

The process can be repeated by determining the worst case price equal to the discounted price P _{n + 1} and by asking the payer for the new price difference δ _{n + 1} to be: ,

U (P _n , 1) = U (P _{n +1} , 0)

This can be repeated until the value P _n reaches the additional discount which the negative value _{(P m + 1) (P} m). That is, the process value (P _m) is more than _{U (P m, 1) =} U (P m + 1, 0) could not be fully discounted enough to be unable to define the value (P _{m + 1)} It can continue until. In this way, the array can be defined to include a plurality of m + 1 prices [P ₀ , P ₁ , ... P _m ] as follows:

[P ₀ , P ₁ , P ₂ ... P _m ] (array 1)

Continuing with the above example for illustrative purposes,

(P ₀ , P ₁ , P ₂ , P ₃ , P ₄ ) = [$ 5000, $ 4500, $ 3800, $ 2800, $ 1500]

(Where m = 4).

Additional data points (e.g., $ 1500), the order to add, the reservation system 16 is value having the value: P _m (P _{m + 1)} (e. G., P ₅ = $ 200) under P _m a Can be limited. The reservation system 16 further restricts P _{m + 1} to the best price so that u _p (P _{m + 1} ) = 1 and associates P _{m + 1} with the worst case value to obtain V _{m + 1} = O May define a travel offer ( _{Tm + 1} ). The payment may be required to limit the minimum value (V _m) which satisfy the following formula 17:

_{U (P m, V m)} = U (P m +1, 0) ( Formula 17)

In one embodiment of the invention, the value (V _m) can be determined by asking a person to pay for one or more parameters that can sufficiently change the travel proposed parameter to be equal to each other, the composite utility of T _m and T _{m + 1.} For a travel offer ( _Tm , _{Tm + 1} ) that includes airline flights and hotels, the payer can either increase the hotel rating from the lowest available class, or reduce the flight duration from the longest available flight, a may be required that the _{(P m, V m) =} U (P m +1, 0) may be sufficient enough to increase the value of V _m to be.

By way of example, the worst case travel offer may have a flight duration of 20 hours and a star rating. In this illustrative case, the payer may respond by raising the hotel rating to four, which may be sufficient to satisfy Equation 17. The payer may be asked to further determine how much the flight duration needs to be reduced to satisfy Equation 17, whether or not the hotel class is below the requested level (e.g., third grade). The payer may be asked how much the flight duration needs to be reduced to satisfy Equation 17 in a travel proposal that maintains a two-star hotel rating. This question may be answered in response to, for example, indicating that the payee may not be sufficient to satisfy Equation 17 simply by raising the hotel rating to the highest rating (e.g., 5 stars).

In any case, utility value (V _m) may take the ratio of the case of the improved parameters and the worst parameters determined by entering the ratio in formula 1. By way of example, the value utility function of the two parameters may have a flight duration weight (w _{c_float duration} = 2/3) and a hotel class weight (w _{c_hotel_class} = 1/3). If the payer responds that if it reduces the flight time from 20 hours to 17 hours, then Equation 17 can be met, then in this example the value (V _m ) can be given by:

(수식 18)

(Eq. 18)

At block 90, the price weight w _p or balance factor may be determined based on the values of m and V _m . From equation (15) it is known that U (P _n , 0) = U (P _n-1 , 1). Substituting these values into Equation 3 yields:

w _p xu _p (P _n ) = w _p xu _p (P _{n -1} ) + (1-w _p )

This can be simplified as follows:

(수식 20)

(Equation 20)

Therefore, if n increases by each increment, u _p (P _n ) increases by (1 - w _p ) / w _p . u _p (P ₀ ) = 0, u _p (P _n ) can be determined for each value of n using the following relationship:

(수식 21)

(Expression 21)

The value of n is as follows:

(수식 22)

(Equation 22)

u _p value (P _{m + 1)} of the (P _{m + 1)} = ₁ in the best case (for example, P ₅ = $ 200) and the relation _{U (P m, V m)} = U (P m + 1, 0 (T _m and T _{m + 1} ) having the worst case value (V _{m + 1} = 0) satisfying the following equation (3), and substituting this price and value utility into Equation 3, Become like this:

w _p = w _p x u _p (P _m ) + (1 - w _p ) x V _m (Equation 23)

Substituting u _p (P _m ) using equation 21 yields:

(수식 24)

(Equation 24)

Solving equation 24 for w _p yields:

(수식 25)

(Equation 25)

Equation 25 can be used to determine the price weight (w _p ) for each value of P _n using the values of m and V _m . Using the values of m and V _m determined for an exemplary travel offer, the value of w _p in price (P _m ) can be determined by substituting these values into equation (25). Using this value in the present example, Equation 25 becomes:

(수식 26)

(Equation 26)

5, a graph 92 having a horizontal axis 94 representing the price P _n of the travel offer T _n and a vertical axis 96 representing the price weight w _p is shown in the numerical example And a graph 98 showing w _p determined for price (P _n ). This graph 98 may define the price utility function u _p (P) in the above example.

For a travel offer with a price between the mapped points, the value of w _p can be determined by connecting the points to the linear segments 100 - 104 and projecting the price on the final graph 98. To provide a smoother, more accurate mapping function, additional information may be derived from the payer to add additional points to the graph 98 between the mapped points P _k and P _{k + 1} . For example, a new point P ₆ representing a price having utility u _p (P ₆ ) = 0.5 may be added between point P ₂ and point P ₃ . The value utility (V _new ) can be determined by equation 27:

(수식 27)

(Equation 27)

Starting at the nearest mapped point (P _k ) (e.g., P ₂ = $ 3800) that has a higher price than the point to be added, the payer may be asked for price P :

U (P, 0) = U (P _k , V _new )

The price P can be mapped to a new utility point between u _p (P _k ) and u _p (P _{k + 1} ).

At block 106 of process 80, reservation system 16 may extract value preferences from the traveler. These value preferences can be used to determine the value utility function (V). To this end, the reservation system 16 may ask the traveler to choose a preferred travel offer at any cost. This question may be answered in response to the traveler having bookmarked the trip, in response to the reservation system 16 displaying the search results, or at any other appropriate time during the travel booking process. For example, after a traveler has booked a travel proposal, the reservation system 16 may ask the traveler which of the plurality of travel proposals satisfying the trip requirement is a travel proposal that can be selected if the price was not an element have. The data representing the traveler's choice can be stored in a database and used to determine the value utility function (V).

Referring now to FIG. 6, the flowchart illustrates a process 110 that may be used to determine the parameter weight set [w _c1 , w _c2 , ... w _cn ] for the value utility function (V) do. The process 110 may be performed, for example, by the reservation system 16 or by any other suitable computer system.

At block 112, the process 110 may select one parameter weight set from a plurality of parameter weight sets. Each parameter set of weights is of at least one of the parameters weight (w _{c_x)} weight parameters including the [w _{c_1, c_2} w, w ... _{c_n]} having a different value than the weight (w _{c_x)} corresponding to other parameters from the set of weights Array. The parameter weight set may provide a range of values for each parameter weight. Each parameter weight may be represented by an equal number of discrete values over a plurality of weight sets, or by a different number of discrete values over a plurality of weight sets. If the number of discrete values varies from one parameter weight to the next, then the number of discrete values may vary depending on the desired sensitivity of the value utility function (V) for the particular parameter. For example, a parameter that has a relatively small impact on the degree to which a traveler values a trip suggestion may have only two possible weights. In contrast, a parameter that has a relatively large impact on the extent to which a traveler tries to value a trip offer may have five or more possible weights. In response to selecting a set of weights, the process 110 may proceed to block 114.

At block 114, the process 110 may select one traveler to analyze from a plurality of travelers in the travel database. The travel database can be maintained, for example, by the GDS and / or the reservation system 16. [ A plurality of travelers can be defined based on common characteristics, such as business travelers traveling for a particular company that traveled for a specified period of time, or can include all travelers in the travel database. In response to selecting the traveler, the process 110 may proceed to block 115.

At block 115, the process 110 may select a trip taken by the traveler. This trip can be identified in the travel database based on the identity of the traveler, the pair of cities at the destination and the origin, the date of the trip, or any other suitable characteristic or combination of characteristics of the trip. The trip nature may include a trip requirement that includes a search term used to search for travel suggestions matching the trip requirement. In any event, a corresponding plurality of travel offers that may have satisfied the trip requirement may be associated with a trip in the travel database, or may be generated in response to the process 110 selecting a trip to analyze. As discussed above, one of the plurality of travel offers may be flagged or otherwise identified with a travel offer preferred by the traveler, regardless of cost. The process 110 may determine the value utility for the preferred travel offer by applying the selected set of parameter weights to the preferred travel offer using Equation 1 before proceeding to block 116. [

At block 116, the process 110 may select a travel offer from a plurality of travel offers before proceeding to block 118. [ At block 118, the process 110 may determine the value utility for the selected travel offer by applying the selected set of parameter weights to the selected travel offer using Equation 1. [ The process may proceed to block 120.

At block 120, the process 110 may determine if the value utility (V) of the preferred travel offer is higher than the value utility (V) of the selected travel offer. In response to the value utility V of the preferred travel offer being higher than the value utility V of the selected travel offer ("Yes" branch at decision block 120), the process 110 proceeds to block 122 And may increment the match score of the selected parameter weight set before proceeding to block 124. In response to the value utility (V) of the preferred travel offer not being higher than the value utility (V) of the selected travel offer ("no" branch at decision block 120) It may proceed to block 124 without incrementing the match score.

At block 124, process 110 may determine whether the currently selected travel offer is the last of the plurality of travel offers associated with the trip being analyzed. That is, the process 110 may determine whether a value utility has been determined for each travel offer in the plurality of travel offers. In response to determining that the currently selected travel offer is not the last travel offer (a "no" branch in decision block 124), the process 110 proceeds to block 126 and returns to block 118 You can choose your next travel offer from multiple travel offers before. In response to the determination that the currently selected travel offer is the last travel offer ("yes" branch at decision block 124), the process 110 may proceed to block 127.

At block 127, process 110 may determine if there are any additional trips taken by the traveler that have not been analyzed. If no further trips taken by the traveler have been analyzed ("no" branch at decision block 127), the process may proceed to block 128. At block 128, the process 110 may select another trip taken by the traveler for analysis. The process 110 may use the current set of parameter weights prior to returning to block 116 to determine the value utility of the preferred travel offer for the selected trip. The updated value utility in the preferred travel offer may be available for comparison with the value utility determined for other travel offers. If there is no additional trip taken by the traveler to be analyzed, the process 110 may proceed to block 129.

At block 129, the process 110 may determine whether the currently selected traveler is the last of the plurality of travelers analyzed. That is, the process 110 determines whether each traveler of a plurality of travelers has been analyzed. If the traveler is not the last traveler in the plurality of travelers being analyzed (the "no" branch in decision block 129), then the process 110 proceeds to block 130 and returns the next traveler You can choose.

If the currently selected traveler is the last traveler of a plurality of travelers ("yes" branch at decision block 129), then the process 110 may proceed to block 132. At block 132, process 110 may determine that the currently selected parameter weight set is the last parameter weight set. If the match score has not been determined for each parameter weight set in the plurality of parameter weight sets ("NO" branch at decision block 132), then process 110 may proceed to block 134. At block 134, the process 110 may select the next set of parameter weights before returning to the block 114.

In response to determining that the currently selected parameter weight set is the last parameter weight set to be analyzed ("Yes" branch at decision block 132), the process 110 may proceed to block 136. At block 136, process 110 may select a set of parameter weights with the highest match score. The set of parameter weights with the highest match score may represent a parameter weight that most likely identifies the travel offer that is preferred to have a higher value utility than other non-preferred travel offers. If more than one parameter weight set has the same match score, the average or multiple value utility weight set may be determined by calculating the barycenter of the parameter weight set with the highest match score. This center of gravity can be determined by any suitable means, such as calculating the centroid of each parameter weight.

In order to prevent a traveler having a large number of business trips from distorting the results, the match score for each parameter weight can be changed to provide equal weight to each traveler's preference. In an alternative embodiment, the match score may include a number of trips taken by a single traveler, who in this case may have a potentially large effect on the final complex parameter weight set. In any case, the composite value utility weight set can be used with Equation 1 to determine the value utility for the travel offer as described above.

Those skilled in the art will appreciate that process 110 shown in FIG. 6 is merely exemplary and that additional variations in process 110 can be used to determine a set of parameter weights. For example, including a process loop may be changed so that multiple parameter weights are scored for each travel offer before selecting the next travel offer. Embodiments of the present invention are not limited to the illustrated process 110.

At block 138 of process 80, process 80 compares the balance factor and internal price preferences extracted from the payer at blocks 88 and 90 and the value preferences extracted from the traveler at block 106 . These aggregated parameters can be used to provide a coefficient to the equation defining the complex utility function U (P, V). In response to receiving the search query from the user system 12, the process 80 may receive utility function coefficients, search result data 140, and traveler personal data 142. The search result data 140 may include information about the price, availability, and other qualifying characteristics of a travel offer that includes goods such as airline flights, hotel rooms, and rental cars. The search results data 140 may be stored in a database associated with merchandise provider system 14, reservation system 16, or any other suitable < RTI ID = 0.0 > Can be obtained from one or more of the databases. The process 80 may be performed on a particular traveler, such as a frequent flyer or member of a frequent club, a personal preference for the seat position of the airplane, a preferred rental car characteristic, Personal data 142 can be received.

At block 144, the process 80 may select a number of travel offers from a plurality of travel offers identified by the search result data 140. These travel offers can be selected based on the composite utility score that can be determined for each travel offer among a plurality of travel offers using Equation 3 as described above. For example, the process 80 may select five trip offers with the highest composite utility score U (P, V) and display the travel offer as the best or most relevant search result to the traveler.

Since the process 110 described in FIG. 6 identifies a single set of weights reflecting the preferences of a plurality of travelers, the tour offer selected at block 144 is based on a single "average" traveler. In other words, the composite utility used to rank travel proposals may not take traveler-specific value preferences into account. However, it may often be desirable to provide a more personal recommendation by selecting a set of weights based, at least in part, on the personal preferences of the particular traveler for whom the travel offer is displayed.

Referring now to FIG. 7, the flow chart illustrates a process 150 that may be used to determine a plurality of "general traveler profiles" for use in classifying travelers. These profiles can be used to group travelers into multiple groups based on one or more similarities in the traveler's preferences. The process 150 may determine a set of parameter weights for each profile based on an optimal number of profiles and an analysis of a plurality of travelers. The plurality of travelers may include, for example, all travelers traveling for a particular business during a specified time window, for example, last year. The traveler who performed the search may be assigned a profile having a parameter weight set that most closely matches his or her personal preference.

The proximity of matching can be determined based on the "distance" between the traveler and the profile. This distance may be based on the number of preferred travel options matched using a set of parameter weights associated with the profile. For example, this distance may be provided by the following distance function:

(1-P _matching ) (Expression 29)

Where P _matching represents the percentage of preferred travel options correctly identified using the set of parameter weights associated with the profile. Thus, a profile having a set of parameter weights that correctly identified all of the preferred travel profiles selected by the traveler can mass-produce the distance of O from the traveler. In contrast, a profile having a set of parameter weights that does not identify any of the preferred travel offers selected by the traveler can mass produce a distance of one from the traveler.

By using a set of parameter weights associated with the closest profile to the traveler, the value utility function (V) used to determine the composite utility value is greater than that of the system using a single parameter weight set across all travelers, Can be better reflected. Through this, the value utility function (V) can more closely match the traveler's preference value preference. Embodiments of the reservation system 16 including a plurality of profiles may provide more personal recommendations than embodiments using a single profile.

At block 152, the process 150 may split a plurality of travelers into a training sub-set and a test sub-set. The traveler in the training sub-set can be used to determine the set of parameter weights to be associated with each profile and the number of profiles to be defined (K). The traveler in the test sub-set can be used to determine the extent to which the defined profile identifies the preferred travel option selected by the traveler in the test sub-set using the training sub-set. The training and test sub-sets may be defined by randomly selecting travelers from a plurality of travelers for inclusion in a training or test sub-set. In one embodiment of the invention, the training sub-set may comprise 70% of a plurality of travelers, and the test sub-set may comprise the remaining 30% of a plurality of travelers. However, those of ordinary skill in the art will appreciate that other ratios may be used to divide the traveler between the training and test sub-sets. Embodiments of the present invention are not limited to the exemplary 70/30 ratio described above. In response to defining the traveler's training and test sub-set, the process 150 may proceed to block 154. [

In block 154, process 150, K ₁ - distance for a set - for each value of the plurality of values of K _M (K) (e.g., K = 2 to K = 10) for the training sub A set of K profiles each having a set of parameter weights that minimize the function (1 - P _matching ) can be determined. This can be accomplished using cluster analysis techniques. For example, the K initial parameter weight sets may be initially randomly selected by associating different sets of parameter weights with each of the K provisional profiles. Each traveler in the training sub-set can be assigned to the closest temporal profile determined by the distance function (1 - P _matching ) to define K subsets of travelers. The new parameter weight set may be selected for each subset of travelers that minimizes the distance function (1 - P _matching ) for this subset. Each traveler in the training sub-set can be reassigned to the closest temporal profile using a new set of parameter weights to compute the distance function (1 - P _matching ). This process can be repeated until each parameter weight set converges to a fixed set of values. Clustering algorithms can be repeated several times using different initial parameter weight sets to prevent convergence to local minimums from affecting the results.

For a given plurality of travelers to be analyzed, an optimal number of profiles to create a minimum amount of travel suggestion selection errors for travelers in the test sub-set by balancing the profile fit verses profile predictive value (K). For example, by defining multiple profiles K, the average distance between travelers in the training sub-set and the closest matching profile can be less than the average distance for a small number of profiles K. That is, if K is large, the average fit value between the matching profile and the traveler can be better. However, as K increases, the number of sample points (e.g., training travelers) available to define each profile in response thereto may be reduced. A profile based on a smaller number of sample points may have a reduced predicted value compared to a profile based on a larger number of sample points so that the final value utility function may be less accurate.

To illustrate that the predicted value decreases as expected when the number of profiles is increased beyond the optimal number of profiles, a penalty to increase the value of K using a regularization parameter (?) Is provided to the training set . The normalization parameter [lambda] may reduce over-fitting by adding a cost value proportional to K. [ This cost value may be added to the distance function provided by equation (29). In one embodiment of the invention, the cost value is a product of the number of normalization parameters (lambda) and the number of profiles K, so that the following error function can be produced:

(1 - P _matching ) +? X K (Equation 30)

Here, (1 - P _matching ) represents the distance between the traveler and the corresponding profile over a plurality of travelers being analyzed, and λ x K represents a cost value associated with an increase in K. The normalization parameter (lambda) provides an adjustable parameter that can be used to balance the predicted and tailored values for a plurality of travelers being analyzed.

If the profile is determined for each value of K, then the process 150 may proceed to block 156. In block 156, the step 150 is a plurality of qualified parameters - may determine the value of K that minimizes the error function for each of the (λ ₁ λ _N). The error function may be computed at block 154 using the traveler in the training sub-set and parameter weight set determined for each profile. In the illustrated embodiment, the normalization parameter set of the present invention gajyeoseo values starting at λ ₁ = ₁ x 10 ^-4 to increase geometrically as the common rate of about 2 to exit from the _{λ 13 = 0.4, λ 1 =} 1 x 10 ^-4 , λ ₂ = 2 × 10 ^-4 , λ ₃ = 4 × 10 ^-4, λ ₁₃ = 0.4. However, other types of sequences, values, and numbers of normalization parameters (?) May be used, and embodiments of the present invention are not limited to numerical examples for K and? Provided herein. In any case, each normalization parameter [lambda] pairs a plurality of normalization parameter / profile number pairs ([lambda] ₁ , K [ _lambda ] ₁ ), ([lambda] ₂ , K _{lambda 2} ) with a corresponding K that produces a minimum error in the training sub- ... it is possible to mass-produce (λ _N, K _λN).

At block 158, the process 150 uses the traveler in the test sub-set to determine the distance provided in equation 29 to each normalization parameter (K _{? 1} , K _{? 2} ... K _{? N} ) The value for the function (1 - P _matching ) can be determined. The profile set associated with the normalization parameter (K _x ) that produces the lowest distance function value in the test sub-set may be selected to identify the optimal profile set.

The process 150 may proceed to block 160 and set the profile in the reservation system 16 according to the optimal set of profiles. A tourist training sub-determining using a set of normalized parameters / profile pair _{((λ 1, K λ1)} , (λ 2, K λ2) ... (λ N, K λN)) , and the tourist test sub- By using the set to determine the optimal normalization parameter / profile pair (? _X , K _x ), embodiments of the present invention allow a traveler to build- in check.

In operation, the traveler can be assigned a profile that provides the best match score between the traveler's preferred travel options and the set of parameter weights associated with the assigned profile, to the optimal profile set. If two profiles produce the same match score, one of the two profiles can be selected at random, or the traveler can be required to select an additional preferred travel option that breaks the tie from multiple travel options. The travel offer may be selected to essentially display the value utility function (V) for the traveler, as described above with respect to Figures 5 and 6, using a set of weight parameters associated with the selected profile. By assigning the traveler to one of the plurality of profiles, the composite utility U (P, V) can select a travel offer that more closely reflects a traveler's preference than an embodiment without a typical traveler profile feature.

Those skilled in the art will appreciate that the program code embodying any of the embodiments of the invention described herein may be distributed separately or collectively to a number of different types of program products will be. In particular, the program code may be distributed using a computer-readable medium and a communication medium, which may include a computer-readable storage medium. A computer-readable storage medium, which is essentially non-volatile, includes volatile and nonvolatile, nonvolatile, nonvolatile, nonvolatile, nonvolatile, nonvolatile, nonvolatile, And removable and non-removable tangible media. The computer-readable storage medium may be RAM, ROM, erasable programmable Readable memory (EPROM), electrically erasable programmable read only memory (EEPROM), flash memory or other solid state memory technology, portable compact disc read only memory (CD-ROM), or other optical storage, magnetic cassettes, Tape, magnetic disk storage or other magnetic storage device, or any other medium that can be used to store the desired information that can be read by the computer. The communication medium may embody computer readable instructions, data structures, or other program modules. By way of example, and not limitation, communication media may include wired media, e.g., wired networks or direct-wired connections, and wireless media such as acoustic, RF, infrared and other wireless media. Any combination of the above may also be included within the scope of computer readable media.

The methods described herein may be implemented by computer program instructions supplied to a processor of any type of computer to produce a machine with a processor executing instructions that implement the functions / operations specified herein. These computer program instructions may be stored in a computer readable medium that can cause the computer to function in a particular manner. To this end, the computer program instructions may be loaded into a computer and executed when executed to provide a process for performing a series of operating steps and creating a computer implemented process to implement the functions / operations specified herein.

Further, the program code described herein may be identified based on an application or software component that implements the program code in a particular embodiment of the present invention. However, it is to be understood that any particular program term is only used for convenience, and therefore the present invention should not be construed as being limited to being used in any particular application that is identified and / or implied by such term. It is further understood that the various features, applications, and devices described herein may be used alone or in combination. Furthermore, it will be appreciated that the manner in which a computer program may be organized into routines, procedures, methods, modules, objects, and / or the like may be applied to general computing systems (e.g., operating systems, libraries, APIs, applications, applets, It is understood that the present invention is not limited to any particular organization and assignment of program functionality described herein, as there are generally no endings to the many ways in which multiple software layers residing on one or more hardware platforms can be allocated.

The terminology used herein is for the purpose of illustrating particular embodiments and is not intended to limit the invention to these embodiments at all. As used herein, the singular terms "a" and "said" are intended to include the plural forms unless the context clearly dictates otherwise. The term " comprising "and / or" comprising "when used herein should be interpreted as referring to the presence of stated features, integers, steps, operations, elements and / , Steps, operations, elements, components, and / or groups are not intended to be < / RTI > Furthermore, the terms "comprising", "having", "having", "having", "consisting", or variations thereof, when used in either the detailed description or the claims, And is intended to be inclusive in a like manner.

According to one aspect, the invention features the following:

1. A computer program product characterized by a plurality of travel offers satisfying a first travel request,

Non-transitory computer readable storage medium; And

Program instructions stored on the non-transitory computer readable storage medium, the program instructions causing the processor to:

Receiving first data identifying a preferred travel offer for each of a plurality of first travelers, wherein each preferred travel offer is selected from the plurality of travel offers;

Analyzing the first data to define a first utility function to quantify the value of each travel offer;

Receiving second data on price sensitivity from a payer;

Defining a second utility function that analyzes the second data to quantify the price of each travel offer; And

Wherein the third utility function comprises an output for quantifying the combined utility for each travel offer, wherein the third utility function is operative to define a third utility function as a weighted sum of the first and second utility functions, product.

2. The method of claim 1, wherein the first utility function comprises a weighted sum of a plurality of fourth utility functions, each travel offer is defined by a parameter set, 4 < / RTI > utility function.

3. The method of claim 2, wherein the program instructions cause the processor to:

The operations defining the plurality of weight sets, each set of weights having different weight combinations, each weight corresponding to one of the fourth utility functions;

Determining a match score for each weight set of the plurality of weight sets based on the first data; And

Selecting the set of weights with the best match score to weight the sum of the fourth utility functions

And to analyze the first data to define the first utility function.

4. The system according to any one of 1 to 3, wherein the program instructions cause the processor to:

Defining a first travel offer with a first price and a first set of parameters;

Defining a second trip offer with a second price and a second set of parameters; And

The third data defining both the first and second prices and all but one of the first and second sets of parameters, and wherein the third data further comprises: 2 data defines one of the first and second price and the first and second parameter sets that are not defined by the third data and the second data is a combination of the first and second prices, And the compound utility for the second travel offer determined by the payer.

5. The method according to any one of 1. to 4. above, wherein the program instructions cause the processor to:

Receiving a second travel request from a second traveler, wherein the second travel request shares a common parameter with the first travel request;

Determining a second plurality of travel offers that satisfy the second travel request;

Ranking the second plurality of travel offers based on an output of the third utility function; And

Select one or more travel offers of the second plurality of travel offers to display on the second traveler based on the ranking.

The embodiments of the present invention are illustrated by way of illustrative example, and these embodiments have been described in considerable detail, but it is not intended that the applicant intend to limit or limit the appended claims to such detail. Additional advantages and modifications will readily occur to those of ordinary skill in the art. The invention is not limited to the specific details, representative apparatus and methods, and illustrative examples shown or described in the broader aspects. Accordingly, it will be apparent to one skilled in the art that the present invention may be practiced without departing from the spirit and scope of the general inventive concept thereof.

Claims (15)

CLAIMS What is claimed is: 1. A method for characterizing a plurality of travel offers satisfying a first travel request,
Each preferred travel offer is selected from a plurality of travel offers, and in the computer, receiving first data identifying a preferred travel offer for each of the plurality of first travelers;
Analyzing the first data to the computer to define a first utility function that quantifies the value of each travel offer;
Receiving second data on price sensitivity from a payer;
Analyzing the second data with the computer to define a second utility function that quantifies the price of each travel offer; And
The third utility function having an output that quantifies the composite utility for each travel offer, the method comprising: defining a third utility function as a weighted sum of the first and second utility functions . The method according to claim 1,
Wherein the first utility function comprises a weighted sum of a plurality of fourth utility functions, each travel offer is defined by a parameter set, and wherein each parameter of the parameter set provides an input to one of the fourth utility functions . ≪ / RTI > 3. The method of claim 2,
Wherein analyzing the first data to the computer to define the first utility function comprises:
Defining a plurality of weight sets, each weight set having a different weight combination, each weight corresponding to one of the fourth utility functions;
Determining a match score for each weight set of the plurality of weight sets based on the first data; And
Selecting a set of weights having a best match score to weight the sum of the fourth utility function. The method of claim 3,
Wherein determining a match score for each weight set comprises:
Defining a weighted sum of a fourth utility function based on the set of weights for which the match score is determined;
Determining, for each first traveler, an output of the first utility function for each travel offer;
Comparing an output of the first utility function to a preferred travel offer with an output of the first utility function for each of the other travel suggestions; And
Each time the output of the first utility function for a preferred travel offer is higher than the output of the first utility function for a travel offer with which the preferred travel offer is compared, the match score of the weight set for which the match score is determined ≪ / RTI > The method according to claim 1,
Defining a first travel offer comprising a first price and a first set of parameters;
Defining a second travel offer comprising a second price and a second set of parameters; And
And providing third data to the payer,
Wherein the third data defines both the first and second price and one of the first and second sets of parameters and the second data defines the first and second parameters not defined by the third data, Set and one of the first and second prices, wherein the second data is such that the compound utility for the first travel offer is the same as the compound utility for the second travel offer determined by the payer, ≪ / RTI > 6. The method of claim 5,
The second data defining the second price, and
Wherein analyzing the second data with the computer comprises:
Setting a sample number equal to a first numerical value;
Defining the first set of parameters to produce a maximum output of the first utility function;
Defining the second set of parameters to produce a minimum output of the first utility function;
Defining the first price to produce a minimum output of the second utility function;
Associating the first price with the sample number in the array; And
In response to the second price being limited,
Incrementing the sample number;
Associating the second price with the sample number in the array;
Setting the first price equal to the second price;
Defining a new second price from the second data;
Setting the second price equal to the new second price; And
Incrementing the sample number as long as the new second price can be limited; associating the second price with the sample number in the array; and comparing the first price with the second price Determining the second price from the second data; and setting the second price equal to the new second price. ≪ Desc / Clms Page number 21 > The method according to claim 6,
Setting the sample size equal to the sample number in response to the inability to define the new second price;
Limiting the first price to be equal to the lowest price in the array;
Defining the second price to produce a maximum output of the second utility function;
Defining the second set of parameters to produce a minimum output of the first utility function;
Determining the first parameter set from the second data such that the first parameter set makes the compound utility for the first travel offer equal to the compound utility for the second travel offer determined by the payer Determining the first parameter set;
Determining a weight for the second utility function based on the output of the first utility function and the sample size for the first parameter set;
Incrementing the sample number;
Associating the second price with the sample number in the array; And
Further comprising: defining an output of the second utility function for each price in the array based on the weight of the second utility function and the sample number corresponding to each price in the array . The method according to claim 1,
The second travel request sharing a common parameter with the first travel request and receiving a second travel request from the second traveler;
Determining a second plurality of travel offers that satisfy the second travel request;
Ranking the second plurality of travel offers with the computer based on an output of a third utility function; And
Selecting one or more travel offers of the second plurality of travel offers to display to a second traveler based on the ranking. The method according to claim 1,
Defining a first sub-set of first travelers from a plurality of first travelers;
The second sub-set includes at least one first traveler not included in the first sub-set; defining a second sub-set of the first traveler from the plurality of first travelers;
Each profile set including a different number of profiles, the profile set defining a first plurality of profile sets;
Determining, for each profile set of the first plurality of profiles, a set of weights for each profile of the profile set that minimizes the distance between the first traveler and the set of profiles in the first sub-set;
Defining a plurality of normalization parameters;
Determining, for each normalization parameter, a profile set of the first plurality of profile sets that minimizes an error function using the first traveler in the first sub-set; and determining the profile set as a second plurality of profile sets ; And
And selecting the profile set from the second plurality of profile sets to minimize the distance function for the second sub-set of the first traveler,
Wherein analyzing the first data to the computer to define the first utility function comprises: associating each first traveler with a profile of the selected set of profiles; and determining a first utility function for the first traveler ≪ / RTI > using the set of weights of the associated profile to define a set of weights for the associated profile. An apparatus characterized by a plurality of travel offers satisfying a first travel request,
A processor; And
A memory operably coupled to the processor,
Wherein the memory comprises a program instruction, and the program instruction, when executed by the processor, causes the device to:
Each preferred travel offer being selected from a plurality of travel offers and receiving first data identifying a preferred travel offer for each of the plurality of first travelers;
Analyzing the first data to define a first utility function that quantifies the value of each travel offer;
Receiving second data on price sensitivity from a payer;
Analyzing the second data to define a second utility function that quantifies the price of each travel offer; And
Wherein the third utility function comprises an output that quantifies a composite utility for each travel offer, the operation further comprising: defining a third utility function as a weighted sum of the first and second utility functions Device. 11. The method of claim 10,
Wherein the first utility function comprises a weighted sum of a plurality of fourth utility functions, each travel offer is defined by a set of parameters, each parameter of the parameter set providing an input to one of the fourth utility functions Lt; / RTI > 12. The method of claim 11,
Program instructions configured to perform the operation of analyzing the first data to define the first utility function,
Each set of weights having a different weight combination, each weight corresponding to one of the fourth utility functions;
Determining a match score for each weight set of the plurality of weight sets based on the first data; And
And selecting the set of weights with the best match score to weight the sum of the fourth utility functions. 11. The method of claim 10,
Wherein the program instructions cause the processor to:
Defining a first travel offer with a first price and a first set of parameters;
Defining a second trip offer with a second price and a second set of parameters; And
And providing third data to the payer, wherein the third data defines both the first and second prices and all but one of the first and second parameter sets, Wherein the data defines one of the first and second price and the first and second set of parameters that are not defined by the third data and wherein the second data indicates that the compound utility for the first travel offer is a payment Wherein the pager is defined by the payer to be a compound utility for the second travel offer determined by the payer. 14. The method of claim 13,
Wherein the program instructions cause the processor to:
Setting the sample number equal to the first numerical value;
Defining the first set of parameters to produce a maximum output of the first utility function;
Defining the second set of parameters to produce a minimum output of the first utility function;
Defining the first price to generate a minimum output of a second utility function;
Associating the first price with the sample number in the array; And
In response to the second price being limited:
Incrementing the sample number;
Associating the second price with the sample number in the array;
Setting the first price equal to the second price;
Defining a new second price from the second data;
Setting the second price equal to a new second price; And
Incrementing the sample number, associating the second price with the sample number in the array, setting the first price equal to the second price, And repeating the operations of: defining an operation, a new second price from the second data, and setting the second price to a new second price. 11. The method of claim 10,
Wherein the program instructions cause the processor to:
The second travel request sharing a common parameter with the first travel request and receiving a second travel request from the second traveler;
Determining a second plurality of travel offers that satisfy the second travel request;
Ranking the second plurality of travel offers based on an output of a third utility function; And
And selecting one or more travel offers of the second plurality of travel offers to display to the second traveler based on the ranking.

Images