KR102249151B1 - Method for arranging goods, computing system and program using the same - Google Patents

Abstract

The product placement method according to an embodiment of the present invention generates child generation product placement sets by applying a genetic algorithm to a plurality of parent generation product placement sets each including product placement information of products to be placed. Step, By performing a local search for sequentially exchanging arrangements between adjacent products among products included in each of the child generation product arrangement sets, the cross-selling rate of the products, the frequency value of cross-selling, and the cross-selling And generating each of the regionally optimized product arrangement sets having a maximum value of the objective function based on the profit value due to and determining an optimal product arrangement based on the generated regionally optimized product arrangement sets.

Description

Product arrangement method, computing system using the same, and program {METHOD FOR ARRANGING GOODS, COMPUTING SYSTEM AND PROGRAM USING THE SAME}

The technical idea of the present invention relates to a product arrangement method, a computing system using the same, and a program, and more particularly, a genetic algorithm and a local search are applied to product arrangement sets including product arrangement information. The present invention relates to a product arrangement method for determining an optimal product arrangement according to the objective function values of product arrangement sets derived by applying together, a computing system using the same, and a program.

Various products are arranged and sold in online/offline stores. The buyer shows a characteristic pattern when purchasing several products, and the seller considers product placement that can maximize product sales by analyzing these patterns.

The cross-selling rate is an index representing the correlation between products seen in purchasing multiple products, and may also mean the probability that two products will be sold together.

In the case of arranging products in consideration of the cross-selling rate, the purchaser may additionally purchase products that were not purchased, although the purchaser had a potential intention to purchase, thereby maximizing product sales.
(Patent Document 1) JP2010-152504 A (2010.07.08)

The task to be achieved by the technical idea of the present invention is optimal according to the value of the objective function of the product placement sets derived by applying a genetic algorithm and a local search together to product placement sets including product placement information. It is to provide a product placement method, a computing system using the same, and a program for determining product placement of the product.

According to the technical idea of the present invention, a product arrangement method according to an aspect of the present invention applies a genetic algorithm to a plurality of parent generation product arrangement sets including product arrangement information of each product to be arranged to place child generation products. Creating sets, by performing a local search for sequentially exchanging arrangements between adjacent products among products included in each of the child generation product arrangement sets, the cross-selling rate of the products, the frequency value of the cross-selling , And the step of generating each of the regionally optimized product arrangement sets, and the objective function value based on the profit value due to cross-selling has a maximum value within the range of the local search, and based on the generated regionally optimized product arrangement sets, the optimal It may include the step of determining product placement.

According to an exemplary embodiment, the product placement method includes a plurality of initial product placement sets including arbitrary product placement information of products to be placed before the step of generating child generation product placement sets by applying the genetic algorithm. By performing the local search for each of the plurality of initial product placement sets generated, the objective function values of the products included in each of the plurality of initial product placement sets are within the range of the local search. It may include the step of generating each of the parent-generation product arrangement sets having a maximum value.

According to an exemplary embodiment, the generating of child generation product placement sets by applying the genetic algorithm comprises: arranging the plurality of parent generation product placement sets according to the objective function value, based on the arranged order. Selecting parent-generation product placement sets to which the genetic algorithm is applied from among a plurality of parent-generation product placement sets, and generating the child generation product placement sets by applying the genetic algorithm to the selected parent-generation product placement sets It may include.

According to an exemplary embodiment, the selecting of the parent generation product placement sets to which the genetic algorithm is applied includes randomly selecting at least three parent generation product placement sets based on the arranged order and including them in each of the candidate groups. And selecting the top two parent-generation product placement sets for each of the candidate groups having the higher objective function value as parent-generation product placement sets to which the genetic algorithm is to be applied.

According to an exemplary embodiment, the step of generating child generation product placement sets by applying the genetic algorithm may include, among the plurality of parent generation product placement sets, the highest 2N (where N is a natural number) the highest objective function value. The parent generation product placement sets may include selecting the child generation product placement sets without applying the genetic algorithm.

According to an exemplary embodiment, the product placement method repeats the step of generating the child generation product placement sets until the number of the child generation product placement sets becomes the same as the number of the plurality of initial product placement sets. I can.

According to an exemplary embodiment, the determining of the optimal product arrangement may include determining the optimal product arrangement based on a region optimized product arrangement set having the largest value of the objective function among the region optimized product arrangement sets. have.

According to an exemplary embodiment, each of the parent generation product arrangement sets may include multidimensional product arrangement information of products to be arranged in a matrix form.

According to an exemplary embodiment, generating the child generation product placement sets by applying the genetic algorithm may include performing a multidimensional cross operation on the selected plurality of parent generation product placement sets.

According to an exemplary embodiment, the multidimensional crossover operation may be an RPMX (Rectangular Partially Matched Crossover) operation.

According to an exemplary embodiment, the RPMX operation may perform a cross operation based on a rectangular cross region.

According to an exemplary embodiment, the crossing ratio in the RPMX operation may be determined according to the size of the objective function value of each of the parent generation product batch sets used in the RPMX operation.

According to an exemplary embodiment, the step of generating the child generation product batch sets by applying the genetic algorithm includes, after the multidimensional cross operation is performed, the parent generation product batch sets on which the multidimensional cross operation is performed. It may further include performing a scramble mutation (scramble mutation).

A computing system according to an aspect according to the technical idea of the present invention applies a genetic algorithm to a plurality of parent-generation product arrangement sets including product arrangement information of each product to be arranged to provide a child generation product arrangement set. Genetic algorithm operation module that generates the product, by performing a local search that sequentially exchanges the arrangements between adjacent products among products included in each of the child generation product arrangement sets, the cross-selling rate and cross-selling of the products The frequency value and the objective function value based on the profit value from cross-selling are based on a local search module that generates each of the regionally optimized product arrangement sets and the generated regionally optimized product arrangement sets having the maximum value within the range of the local search. Thus, it may include a product placement module that determines the optimal product placement.

A program stored in a medium for performing a product placement method combined with a processor according to an aspect according to the technical idea of the present invention is a plurality of parent-generation product placement sets each including product placement information of the products to be placed. Generating child generation product placement sets by applying a genetic algorithm to each field, local search for sequentially exchanging placements between adjacent products among products included in each of the child generation product placement sets By performing the cross-selling rate of the products, the frequency value of the cross-selling, and the step of generating each of the regional optimized product placement sets having a maximum value within the range of the local search objective function value based on the profit value due to the cross-selling. And a program code for performing the step of determining an optimal product arrangement based on the generated regional optimized product arrangement sets.

The method and apparatus according to an embodiment of the present invention determine the optimal product arrangement according to the value of the objective function of the product arrangement sets derived by applying a genetic algorithm and a local search to the product arrangement sets together. By doing so, it has the effect of finding a product arrangement that can maximize product sales.

In addition, the method and apparatus according to an embodiment of the present invention uses an objective function that considers the frequency of cross-selling and profit through cross-selling, thereby maximizing the sales profit of product sales beyond the maximization of product sales. There is an effect that can be found.

A brief description of each drawing is provided in order to more fully understand the drawings cited in the detailed description of the present invention.
1 is a block diagram of a computing system according to an embodiment of the present invention.
2 is a flowchart of a product arrangement method according to an embodiment of the present invention.
3 is a diagram showing an initial product arrangement set according to an embodiment of the present invention.
FIG. 4 is a diagram illustrating an example of a parent generation product placement set generated by performing a local search based on the initial product placement set of FIG. 3.
5 and 6 are diagrams illustrating a process of generating child generation product placement sets by applying a genetic algorithm to the parent generation product placement sets of FIG. 4.

The technical idea of the present invention is that various changes may be made and various embodiments may be provided, and specific embodiments will be illustrated in the drawings and will be described in detail through a detailed description. However, this is not intended to limit the technical idea of the present invention to a specific embodiment, it should be understood to include all changes, equivalents, and substitutes included in the scope of the technical idea of the present invention.

In describing the technical idea of the present invention, if it is determined that a detailed description of a related known technology may unnecessarily obscure the subject matter of the present invention, a detailed description thereof will be omitted. In addition, numbers (eg, first, second, etc.) used in the description of the present specification are merely identification symbols for distinguishing one component from other components.

In addition, in the present specification, when one component is referred to as "connected" or "connected" to another component, the one component may be directly connected or directly connected to the other component, but specially It should be understood that as long as there is no opposite substrate, it may be connected or may be connected via another component in the middle.

In addition, terms such as "~ unit", "~ group", "~ character", and "~ module" described in the present specification mean a unit that processes at least one function or operation, which is a processor or a microcomputer. Processor (Micro Processer), Micro Controller, CPU (Central Processing Unit), GPU (Graphics Processing Unit), APU (Accelerate Processor Unit), DSP (Drive Signal Processor), ASIC (Application Specific Integrated Circuit), FPGA It may be implemented in hardware or software such as (Field Programmable Gate Array), or a combination of hardware and software, and may be implemented in a form combined with a memory that stores data necessary for processing at least one function or operation. .

In addition, it is intended to clarify that the division of the constituent parts in the present specification is only divided by the main function that each constituent part is responsible for. That is, two or more constituent parts to be described below may be combined into one constituent part, or one constituent part may be divided into two or more for each more subdivided function. In addition, each of the constituent units to be described below may additionally perform some or all of the functions of other constituent units in addition to its own main function, and some of the main functions of each constituent unit are different. It goes without saying that it can also be performed exclusively by.

Hereinafter, embodiments according to the technical idea of the present invention will be described in detail in order.

1 is a block diagram of a computing system according to an embodiment of the present invention.

1, a computing system 100 according to an embodiment of the present invention includes an initial product placement set generation module 110, a genetic algorithm calculation module 120, a local search module 130, and an objective function calculation module ( 140), and a memory 150.

The computing system 100 may perform an operation to determine an optimal product arrangement when placing products in an on/offline store for selling products.

According to an embodiment, a data processing process in the computing system 100 may be performed based on hardware included in the computing system 100, for example, a memory 150 and a processor (not shown). In this case, the initial product placement set generation module 110, the genetic algorithm calculation module 120, the local search module 130, and the objective function calculation module 140 may be implemented as some functions of the processor. For example, the product arrangement method according to the embodiment of the present invention may be implemented as a program code and stored in the memory 150, and the memory 150 is combined with the processor to perform the product arrangement method according to an embodiment of the present invention. I can make it.

The initial product arrangement set generation module 110 may generate a plurality of initial product arrangement sets including arbitrary product arrangement information of products to be arranged.

According to an exemplary embodiment, the initial product placement set generation module 110 may receive information on products to be disposed from a user input or the memory 150. In this case, the initial product arrangement set generation module 110 may generate a plurality of initial product arrangement sets by randomly arranging information on the delivered products in a multidimensional space to be arranged. The generated plurality of initial product arrangement sets may be input to the genetic algorithm calculation module 120 as parent generation product arrangement sets.

According to another embodiment, the initial product placement set generation module 110 transfers a plurality of initial product placement sets to the local search module 130, and the local search module 130 is input to the genetic algorithm calculation module 120. A plurality of parent generation product placement sets can be created. In this case, the local search module 130 performs a local search for each of the plurality of initial product arrangement sets, so that the objective function values of the products included in each of the plurality of initial product arrangement sets are the maximum values within the local search range. It is possible to create a plurality of parent-generation product placement sets having a.

The genetic algorithm calculation module 120 may generate child generation product placement sets by applying a genetic algorithm to a plurality of parent generation product placement sets each including product placement information of products to be placed.

According to an embodiment, the plurality of parent-generation product placement sets may be the same as the initial product placement sets generated by the initial product placement set generation module 110.

According to another embodiment, the plurality of parent-generation product placement sets may be product placement sets based on a result of a local search for each of the initial product placement sets by the local search module 130.

The genetic algorithm operation module 120 may include a selection unit 122, a cross operation unit 124, and a mutation operation unit 126.

The selection unit 122 may select parent generation product placement sets to which the genetic algorithm is applied in a corresponding round from among the input plurality of parent generation product placement sets.

According to an embodiment, the selection unit 122 may select at least two parent generation product placement sets to which a genetic algorithm is applied in a corresponding round from among a plurality of parent generation product placement sets.

According to an embodiment, the selection unit 122 arranges a plurality of parent household product placement sets according to an objective function value calculated by the objective function calculation module 140, and arranges a plurality of parent household products based on the arranged order. You can select sets.

According to another embodiment, the selection unit 122 may randomly select at least three or more parent-generation product placement sets and include them in each of the candidate groups. The selection unit 122 may select the parent generation product placement sets to which the genetic algorithm is applied in the corresponding round as the top two parent generation product placement sets having a large objective function value for each candidate group.

According to another embodiment, the selection unit 122 is the top 2N (where N is a natural number) parent-generation products having a large objective function value calculated by the objective function calculation module 140 among a plurality of parent-generation product placement sets. Batch sets can be selected as child generation product batch sets without the application of genetic algorithms.

The cross operation unit 124 may perform a cross operation on the parent generation product placement sets selected by the selection unit 122.

According to an embodiment, when each of the parent generation product placement sets includes product placement information for a multidimensional space, the cross operation unit 124 may perform a multidimensional cross operation.

For example, when each of the parent generation product placement sets includes product placement information in a two-dimensional space, and each of the parent generation product placement sets is configured in a 2D matrix form, the multidimensional crossover operation is RPMX (Rectangular Partially Matched Crossover). It can be an operation. A detailed process of the RPMX operation will be described later with reference to FIGS. 5 and 6.

The mutation operation unit 126 may perform a mutation on each of the parent generation product batch sets cross-calculated by the cross operation unit 124.

According to an embodiment, the mutation operation unit 126 may perform a scramble mutation on each of the parent generation product batch sets that are cross-calculated by the cross operation unit 124.

The genetic algorithm operation module 120 includes a genetic algorithm including at least one of a selection process of the selection unit 122, a cross operation process of the cross operation unit 124, and a mutation operation process of the mutation operation unit 126. Applicable to parent generation product placement sets.

The genetic algorithm is applied to the product batch sets of the parent generation inputted to the genetic algorithm operation module 120, and as a result, product batch sets of the child generation may be generated. The generated child generation product placement sets may be input as parent generation product placement sets back to the genetic algorithm operation module 120 after undergoing a local search process by the local search module 130.

The number of generations of child generation product placement sets may increase according to the number of times the genetic algorithm is applied by the genetic algorithm calculation module 120.

For example, if the number of generations of the first parent-generation product placement sets is 1 generation, the number of generations of its child generation product placement sets becomes 2 generations, and when the genetic algorithm is applied again to the product placement sets of the 2nd generation, the number of generations is 3 generations. Can be.

According to an embodiment, the genetic algorithm operation module 120 may repeatedly apply the genetic algorithm until a product placement set of child generations having a target number of generations preset by the user is generated.

The local search module 130 may perform a local search for sequentially exchanging arrangements between adjacent products among products included in an input initial product arrangement set or a child generation product arrangement set.

A process of performing the local search will be described later with reference to FIG. 3.

In the process of performing the local search, the local search module 130 determines the value of the objective function based on the cross-selling rate, the frequency of cross-selling, and the profit value due to the cross-selling of products included in the product placement set that is the target of the local search. It is possible to create a regional optimized product placement set having a maximum value within the search range.

According to an embodiment, the local search module 130 may determine whether an objective function value increases by exchanging two positions of products included in a product arrangement set to be a target of the local search. As a result of the judgment, when the value of the objective function increases, the exchange is confirmed, and when the value of the objective function is maintained or decreased, the location of the original products is repeatedly performed, thereby optimizing the region where the value of the objective function has the maximum value within the local search range. You can create product placement sets.

The objective function calculation module 140 may calculate and provide an objective function value for the product batch set when it is necessary to calculate the objective function value during the calculation of the genetic algorithm calculation module 120 or the local search module 130. have.

Depending on the embodiment, the value of the objective function f may be determined according to the following [Equation 1].

[Equation 1]

f=

i = 1, ..., nSubject to

i = 1, ..., n

k= 1, ..., n,

k= 1, ..., n,

In the above [Equation 1], i and j denote different products, k and l denote different locations, and weight_profit(Ii,Ij) is the profit value from cross-selling of i and j products The weight based on, weight_frequency(Ii,Ij) may be a weight based on the frequency of cross-selling of i-products and j-products, and lift(Ii→Ij) may mean the cross-selling rate of i-products and j-products.

xik may mean a variable having a value of 1 when product i is located at the k-th position, and having a value of 0 when product i is not located at the k-th position.

xjl may mean a variable having a value of 1 if the j product is located in the l-th position, and may mean a variable having a value of 0 if the j product is not in the l-th position.

y is a variable indicating whether two positions are adjacent, and ykl has a value of 1 if the k-th position and the l-th position are adjacent, and may have a value of 0 if they are not. A case in which the two positions are adjacent in the horizontal direction and a case in which the two positions are adjacent in the vertical direction may be determined as adjacent.

lift(Ii→Ij) can be determined according to the following [Equation 2].

[Equation 2]

That is, the value of the objective function f may be determined based on a cross-selling rate of products arranged adjacently, a frequency value of cross-selling, and a profit value due to cross-selling.

According to an embodiment, weight_profit(Ii,Ij) may be a profit value itself due to cross-selling of product i and product j.

According to another embodiment, weight_profit(Ii,Ij) may be determined based on a profit value resulting from cross-selling of product i and product j in a predetermined weight interval.

Depending on the embodiment, the weight_frequency(Ii,Ij) may be the number of times that product i and product j are cross-selled within a certain period of time or a reciprocal of a cycle in which product i and product j are cross-selled.

According to another embodiment, the weight_frequency(Ii,Ij) may be determined based on the number of cross-selling of products i and j within a certain period in a predetermined weight interval or the reciprocal of a period in which cross-selling of products i and j is performed. .

The memory 150 includes data required for calculation of the initial product placement set generation module 110, the genetic algorithm calculation module 120, the local search module 130, and the objective function calculation module 140 of the computing system 100, or Data according to the calculation result can be saved.

According to an embodiment, the memory 150 may store a program for performing a product arrangement method according to an embodiment of the present invention.

2 is a flowchart of a product arrangement method according to an embodiment of the present invention. 3 is a diagram showing an initial product arrangement set according to an embodiment of the present invention. FIG. 4 is a diagram illustrating an example of a parent generation product placement set generated by performing a local search based on the initial product placement set of FIG. 3. 5 and 6 are diagrams illustrating a process of generating child generation product placement sets by applying a genetic algorithm to the parent generation product placement sets of FIG. 4.

2 to 6, the initial product arrangement set generation module 110 of the computing system 100 may generate a plurality of initial product arrangement sets including arbitrary product arrangement information of products to be placed ( S10).

Referring to FIG. 3, according to an embodiment, each of the initial product arrangement sets may randomly arrange products to be arranged in a multidimensional space, and may include information on the arranged products.

According to an embodiment, each of the initial product arrangement sets may include information on products randomly arranged in a two-dimensional space in a matrix form.

In the matrix illustrated in FIG. 3, a number may mean a number for identifying a product to be disposed, and a location in which the number is written may mean a location where a corresponding product is disposed.

The initial product placement set shown in FIG. 3 illustrates a case in which 25 products are arranged. For example, product 1 is placed at the top left in a square two-dimensional space in which the product is to be placed, and product 5 is a product. It may mean that it is arranged at the top right in the square two-dimensional space to be arranged.

The initial product arrangement set generation module 110 of the computing system 100 may generate an initial product arrangement set as shown in FIG. 3 by a predetermined number by the user.

The computing system 100 may generate parent-generation product placement sets based on the initial product placement sets (S20).

According to an embodiment, the initial product placement set generation module 110 transfers a plurality of initial product placement sets to the local search module 130, and the local search module 130 is a plurality of items to be input to the genetic algorithm calculation module 120. You can create product placement sets of parent generation. In this case, the local search module 130 performs a local search for each of the plurality of initial product arrangement sets, so that the objective function values of the products included in each of the plurality of initial product arrangement sets are the maximum values within the local search range. It is possible to create a plurality of parent-generation product placement sets having a.

Depending on the embodiment, step S20 may be omitted, and in this case, a plurality of initial product batch sets may be input to the genetic algorithm calculation module 120 as parent generation product batch sets.

Referring to FIG. 4, there is shown an example of a parent generation product placement set generated by performing a local search on the initial product placement set shown in FIG. 3.

According to an embodiment, the local search module 130 may determine whether the objective function value of the initial product arrangement set increases after changing positions between products in the initial product arrangement set illustrated in FIG. 3. As a result of the determination, if the value of the objective function increases, it is determined that the positions between the products are changed, and if the value of the objective function is maintained or decreases, the positions of the products can be relocated.

For example, if the objective function value increases after changing the positions of product 1 in the first position and product 18 in the second position, the changed position can be determined, placing product 18 in the first position, and placing product 1 in the second position. have. Next, after changing the positions of the product No. 18 in the second position and the product No. 6 in the third position, you can change the position depending on whether the value of the objective function increases. In the same way, the product at the 24th position and the product at the 25th position can be changed. The local search can be repeated until the location is exchanged.

Depending on the embodiment, a local search is performed sequentially for a product in the first location and a product in the remaining location, and then, a product in the second location and a product in the remaining location sequentially. Also, the method and order of the local search can be variously modified.

As a result of the local search of the initial product arrangement set of FIG. 3, the parent generation product arrangement set of FIG. 4 may be generated.

According to an embodiment, each of the parent generation product placement sets may be generated by performing a local search for each of a plurality of initial product placement sets generated by the initial product placement set generation module 110 of the computing system 100. have.

The genetic algorithm operation module 120 of the computing system 100 may apply a genetic algorithm to the product placement sets of the parent generation to generate the product placement sets of the child generation (S30).

According to an embodiment, the genetic algorithm operation module 120 includes at least one of a selection process by the selection unit 122, a cross operation process by the cross operation unit 124, and a mutation operation process by the mutation operation unit 126. The genetic algorithm can be applied to the product placement sets of the parent generation.

Depending on the embodiment, the cross operation process included in the genetic algorithm may be performed by the rectangular partially matched crossover (RPMX) operation shown in FIGS. 5 and 6.

5 and 6, it is exemplified a case in which a genetic algorithm is applied to two parent generation product placement sets PR1 and PR2 to generate two child generation product placement sets OR1 and OR2.

In each of the two parent-generation product placement sets PR1 and PR2, there are cross-sections RG1 and RG2 that are set to perform RPMX operation. The crossing regions RG1 and RG2 may mean regions that are exchanged with each other as the RPMX operation is performed.

Referring to FIG. 5, the entire cross-region RG1 of the first parent-generation product arrangement set PR1 may be exchanged with the entire cross-region RG2 of the second parent-generation product arrangement set PR2.

Looking at the first parent-generation product placement set (PR1') and the second parent-generation product placement set (PR2') after the cross-regions (RG1, RG2) are exchanged with each other, due to the exchange of the cross-regions (RG1, RG2) There are cases where products are arranged overlapping, for example, 23, 3, 14, and 25 in the case of the first parent-generation product placement set PR1' in which the cross-regions RG1 and RG2 are exchanged with each other. The product overlaps with the product in the exchange area RG2 in which the product was exchanged. In addition, in the case of the second parent-generation product placement set (PR2') in which the crossing areas (RG1, RG2) are exchanged with each other, the products in the crossing area (RG1) in which the 4, 7, 12, and 9 products are exchanged. Overlaps with

In the product arrangement method according to an embodiment of the present invention, since it is assumed that the same product is arranged at one location in the product arrangement set, the overlapping product may be replaced with another product.

Referring to FIG. 6, among products in the remaining areas except for the intersection areas RG1 and RG2 of the parent-generation product placement sets PR1 and PR2, the products that do not overlap even after the exchange of the cross areas RG1 and RG2 are located. I can keep it. For example, for products #15, #2, #20, #22, #24, #21, #19, #11, #13, #16, #6, and #17, the first intersection area (RG1) is the second After being exchanged with the cross-region RG2, the position can be maintained since it does not overlap with the products 14, 10, 18, 3, 8, 1, 23, 5, and 25.

Among the products in the rest of the products except for the intersection areas (RG1, RG2) of the parent-generation product placement sets (PR1, PR2), the locations of products that overlap with the products in the exchange area after the exchange of the intersection areas (RG1, RG2) are It can be replaced with another product.

Referring to FIG. 6, in the case of product No. 14 overlapping according to the exchange of the cross-area (RG1, RG2), find the location of the product that is duplicated in step ① (for example, another product No. 14) and step ② to another parent household. After searching for the product in the location corresponding to the location of the duplicated product in the product placement set (PR2') (e.g., row 3, column 2), the product that was first duplicated as the product searched in step ③ (e.g., product 4) , Product 14) can be substituted.

In this case, when the replaced product overlaps with the product in the exchanged area again, the replacement process of steps ① to ③ may be repeatedly performed.

Through this process, after the product location replacement process in the remaining areas except for the intersection areas RG1 and RG2 and the intersection areas RG1 and RG2 is completed, the two child generation product placement sets OR1 and OR2 Can be created.

The above illustrates a case in which all of the crossing regions RG1 and RG2 are exchanged with each other in the RPMX operation, but the cross ratio of the crossing regions RG1 and RG2 may vary according to embodiments.

According to an embodiment, the intersection ratio of the intersection regions RG1 and RG2 may be determined based on the objective function value of the product arrangement set including the corresponding intersection regions RG1 and RG2.

For example, if the objective function value of the first parent-generation product arrangement set PR1 is '100' and the objective function value of the second parent-generation product arrangement set PR2 is '50', the first intersection area RG1 and The second intersection region RG2 may be exchanged at an intersection ratio of 2:1.

Returning to FIG. 2, the local search module 130 of the computing system 100 may perform a local search on each of the generated child generation product arrangement sets to generate regional optimized product arrangement sets (S40).

According to an embodiment, the local search module 130 exchanges the location of products in each of the generated child generation product placement sets, and an objective function based on the cross-selling rate of products, the frequency of cross-selling, and the profit value due to cross-selling. It is possible to create regionally optimized product placement sets having product placements whose values have a maximum value within the local search range.

The genetic algorithm calculation module 120 or the local search module 130 of the computing system 100 may determine whether the number of generations of the generated regional optimized product placement sets has reached the target number of generations (S50).

Depending on the embodiment, the number of generations may be determined according to the number of times the genetic algorithm is applied by the genetic algorithm calculation module 120, and step S30 of applying the genetic algorithm until a predetermined target number of generations is reached and step S40 of performing a local search The steps can be performed repeatedly.

When the number of households of the regional optimized product arrangement sets generated in step S40 reaches the target number of households, the optimal product arrangement may be determined based on the regional optimized product arrangement sets of the target number of households (S60).

Depending on the embodiment, the number of regional optimized product placement sets generated through step S40 may be plural, and in step S60, the objective function value of each of the plurality of regional optimized product placement sets reaching the target number of households is calculated and the maximum A regional optimized product arrangement set having an objective function value can be determined as an optimal product arrangement.

In the above, the technical idea of the present invention has been described in detail with reference to various embodiments, but the technical idea of the present invention is not limited to the above embodiments, and those of ordinary skill in the art within the scope of the technical idea of the present invention Various modifications and changes are possible by this.

100: computing system
110: Initial product placement set creation module
120: genetic algorithm operation module
130: local search module
140: objective function calculation module
150: memory

Claims (15)

Generating, by the computing system, a genetic algorithm to a plurality of parent-generation product arrangement sets including product arrangement information of products to be arranged by the computing system to generate child generation product arrangement sets;
The computing system performs a local search for sequentially exchanging arrangements between adjacent products among products included in each of the child generation product arrangement sets, thereby cross-selling rates and crossovers between adjacent products among the products. Generating each of the regionally optimized product arrangement sets in which an objective function value based on a frequency value of sales and a profit value due to cross-selling has a maximum value within a range of a local search; And
Determining, by the computing system, an optimal product arrangement based on the generated regional optimized product arrangement sets,
The computing system, when the regional optimized product arrangement sets do not reach a target number of households, inputting the regional optimized product arrangement sets as the parent generation product arrangement sets,
The cross-selling rate is a probability that the adjacent products will be sold together, and the frequency value of the cross-selling is the number of cross-selling of the adjacent products within a certain period,
The local search, in the child generation product placement set, exchanges the positions of products included in the child generation product placement set that is the target of the local discovery among the child generation product placement sets.
The method of claim 1,
The above product placement method,
Before the step of generating child generation product placement sets by applying the genetic algorithm,
Generating, by the computing system, a plurality of initial product arrangement sets including arbitrary product arrangement information of products to be placed; And
The computing system performs the local search for each of the generated initial product arrangement sets, so that the objective function values of the products included in each of the plurality of initial product arrangement sets are within the range of the local search. And generating each of the parent-generation product placement sets having a maximum value.
The method of claim 2,
The step of generating child generation product placement sets by applying the genetic algorithm,
Arranging, by the computing system, the plurality of parent-generation product placement sets according to the objective function value;
Selecting, by the computing system, from among the plurality of parent generation product placement sets to which the genetic algorithm is applied, based on the arranged order; And
And generating, by the computing system, the child generation product placement sets by applying the genetic algorithm to the selected parent generation product placement sets.
The method of claim 3,
The step of selecting the product placement sets of the parent generation to which the genetic algorithm is applied,
Randomly selecting at least three or more parent-generation product placement sets based on the arranged order and including them in each of the candidate groups;
And selecting, by the computing system, top two parent generation product placement sets for each of the candidate groups, with the higher objective function value as parent generation product placement sets to which the genetic algorithm is to be applied.
The method of claim 4,
The step of generating child generation product placement sets by applying the genetic algorithm,
In the computing system, among the plurality of parent-generation product placement sets, the top 2N (the N is a natural number) parent-generation product placement sets having the largest objective function values are the child-generation product placement sets without the application of the genetic algorithm. Including the step of selecting, the product placement method.
The method of claim 5,
The above product placement method,
The computing system repeats the step of generating the child generation product placement sets until the number of the child generation product placement sets becomes the same as the number of the plurality of initial product placement sets.
The method of claim 6,
The step of determining the optimal product arrangement,
And determining the optimal product arrangement based on a region optimized product arrangement set having the largest value of the objective function among the region optimized product arrangement sets.
The method of claim 7,
Each of the parent generation product placement sets,
A product arrangement method including multidimensional product arrangement information of products to be arranged in a matrix form.
The method of claim 8,
Generating the child generation product placement sets by applying the genetic algorithm,
And performing a multidimensional cross operation on the selected plurality of parent-generation product placement sets.
The method of claim 9,
The multidimensional crossover operation,
RPMX (Rectangular Partially Matched Crossover) operation, a product placement method.
The method of claim 10,
The RPMX calculation,
A product placement method that performs an intersection operation based on the intersection area of a rectangle.
The method of claim 11,
The crossing ratio in the RPMX calculation is determined according to the size of the objective function value of each of the parent generation product placement sets used in the RPMX calculation.
The method of claim 12,
Generating the child generation product placement sets by applying the genetic algorithm,
After the multi-dimensional cross operation is performed, the step of performing a scramble mutation (scramble mutation) on the parent-generation product batch sets in which the multi-dimensional cross operation has been performed.
A genetic algorithm calculation module for generating child generation product placement sets by applying a genetic algorithm to a plurality of parent generation product placement sets each including product placement information of products to be placed;
By performing a local search for sequentially exchanging arrangements between adjacent products among products included in each of the child generation product arrangement sets, the cross-selling rate between adjacent products among the products, the frequency value of cross-selling, And a local search module for generating each of the local optimized product arrangement sets in which an objective function value based on a profit value due to cross-selling has a maximum value within a range of the local search. And
When the local optimized product placement sets do not reach the target number of households, inputting the local optimized product placement sets to the genetic algorithm calculation module as the parent household product placement sets,
Based on the generated regional optimized product placement sets, comprising a product placement module for determining an optimal product placement,
The cross-selling rate is a probability that the adjacent products will be sold together, and the frequency value of the cross-selling is the number of cross-selling of the adjacent products within a certain period,
The local search, in the child generation product arrangement set, exchanges the positions of products included in the child generation product arrangement set that is the target of the local search among the child generation product arrangement sets.
As a program stored in a medium for performing a product placement method combined with a processor,
Generating child generation product placement sets by applying a genetic algorithm to a plurality of parent generation product placement sets each including product placement information of products to be placed;
By performing a local search for sequentially exchanging arrangements between adjacent products among products included in each of the child generation product arrangement sets, the cross-selling rate between adjacent products among the products, the frequency value of cross-selling, And generating each of the regional optimized product arrangement sets in which an objective function value based on a profit value due to cross-selling has a maximum value within a range of the local search.
When the regional optimized product arrangement sets do not reach a target number of households, inputting the regional optimized product arrangement sets as the parent generation product arrangement sets; And
Based on the generated regional optimized product arrangement sets, comprising a program code for performing the step of determining an optimal product arrangement,
The cross-selling rate is a probability that the adjacent products will be sold together, and the frequency value of the cross-selling is the number of cross-selling of the adjacent products within a certain period,
The local search, a program for exchanging the positions of products included in the product arrangement set of the child generation that is the target of the local search among the product arrangement sets of the child generation within the product arrangement set of the child generation.

Images