An interactive genetic algorithm-based furniture individualized customization system and method

Abstract

The present application relates to a kind of furniture individualization customization method based on interactive genetic algorithm, to help small space dweller reasonable planning furniture storage space layout, mainly with wardrobe storage space layout.The technical scheme and its main use include the following points:1) the internal storage function module of wardrobe is divided and the space size range of each function module is summarized, and the necessary function module is selected based on Carnot model combined with "Better-worse satisfaction coefficient";2) using interactive genetic algorithm, the optimal storage layout scheme is obtained for user visualization, solves the problem that user preference is fuzzy and communication between designer is high cost and low efficiency;3) introduce penalty function to guide scheme generation direction, to obtain the final optimization result with less evaluation times, relieve fatigue generated by user when evaluating scheme, while ensuring that storage layout scheme is feasible and reasonable.

Description

Technical Field

[0001] This invention relates to the field of furniture personalization, and in particular to a furniture personalization system and method based on an interactive genetic algorithm. Background Technology

[0002] More and more people are choosing to live in small apartments or lofts. Their needs for furniture are no longer limited to the single function of traditional furniture. They require an understanding of their home environment, lifestyle, and storage needs to design personalized and diverse space storage layouts. Customized furniture is increasingly becoming an indispensable part of modern home decoration, offering features such as custom-made options, environmental friendliness, style, and professionalism. However, current personalized customization solutions are often limited to aesthetics. There are no unified standards for storage space layout in the market, relying heavily on the experience of users and designers. This approach not only leads to a disconnect between the finished furniture and the user's actual needs but also increases the human and material costs for both.

[0003] Interactive genetic algorithms, as a type of computer intelligent algorithm that expands the application areas of traditional genetic algorithms, are currently mainly used in design fields such as architecture, music, industry, and clothing. They embed subjective factors such as human consciousness, emotions, and preferences—which cannot be edited into specific objective functions—as important indicators in the evolutionary process into the algorithm system to guide the direction of population evolution. However, there has been no research on the application of interactive genetic algorithms in the design of furniture storage functional layouts. Therefore, the introduction of interactive genetic algorithms into the design of furniture storage functional layouts is of pioneering significance. Summary of the Invention

[0004] The purpose of this invention is to propose a furniture personalization system and method based on an interactive genetic algorithm to solve the above-mentioned problems.

[0005] This invention achieves the above objectives through the following technical solution: a furniture personalization customization system and method based on an interactive genetic algorithm, characterized by comprising the following steps:

[0006] 1) Divide the designated furniture functional modules according to the A1 standard;

[0007] 2) Define the regions for different functional modules and set their size ranges;

[0008] 3) A mathematical model is built for the functional module area using the grid method, and the letter symbol method is used to label it sequentially from left to right along the vertical axis using a to y;

[0009] 4) Set the unit modulus, growth direction and hierarchical variables of the furniture module itself as a gene locus of the furniture chromosome, respectively, to control the phenotype of each functional area of the furniture itself;

[0010] 5) Follow the gene decoding rules to decode all the gene information contained in the chromosome in the order of a~y from small to large to the furniture internal storage cell in turn until all the gene information is decoded, and the entire furniture storage space layout is covered by the dominant phenotype;

[0011] 6) Set four different functional area restriction rules to increase the penalty term of the objective function;

[0012] 7) Select the initial population size as M=60P55, obtain the initial population P54, and perform fitness evaluation, select M=60 as the initial population size, and perform fitness evaluation on the initial population.

[0013] 8) Use single-point crossover method to perform gene crossover and mutation operation, and set the population crossover rate and mutation rate as Pc=0.8, Pm=0.1;

[0014] 9) Output the optimal scheme or the visual furniture scheme when the number of user evaluations reaches forty-five times.

[0015] As a preferred embodiment of the present application, the A1 standard is the optimal standard, and the optimal standard is to divide the furniture into multiple functional areas and set the storage content and commonly used accessories of each functional area.

[0016] As a preferred embodiment of the present application, the interior space of the furniture is divided into five areas (A-E): the A area is to set the space height in the range of 600-1400 mm; the B area is to set the space height in the range of 1400-1800 mm; the C area is to set the space height in the range of above 1800 mm; the D and E areas are to set the space height in the range of below 600 mm.

[0017] As a preferred embodiment of the present application, the corresponding space size range is set according to different functional modules to meet the use scene of the items in the area in actual life.

[0018] As a preferred embodiment of the present application, the cell is numbered, and the code is n, wherein the n cell n=i[a≤i≤y] is the functional module phenotype gene information of the initial growth point, which is represented in the form of a three-element array (Ai, Bi, Ci), wherein A represents the unit modulus variable, B represents the growth direction variable, and C represents the hierarchical variable.

[0019] As the preferred embodiment of the present application, the length of the complete chromosome genotype is 25*3=75 bits, and the acquisition method is to sequentially connect the gene bit information in the a-y cells in order.

[0020] As the preferred embodiment of the present application, the phenotypes of the eight functional modules are expressed in the form of a ternary array in the chromosome individual genes, and the functional module translation rules are made into a table.

[0021] As the preferred embodiment of the present application, the phenotypes are divided into recessive or dominant phenotypes, the recessive phenotype is hidden and invisible in the process of user participation in interactive genetic algorithm, and the dominant phenotype is visible; the decoding priority is used as the basis for the recessive or dominant phenotype of the eight modules, and the earlier the cell letter number in the English letter sequence table, the more preferred the decoding order; here, only the overall furniture storage layout scheme generated by the dominant phenotype is presented to the user for scoring in a visual effect.

[0022] As the preferred embodiment of the present application, a penalty function is introduced, and a penalty mechanism is developed, except for the selected dominant phenotype, the fitness of the remaining same dominant phenotype chromosome individuals is set to 0 and hidden, avoiding participation in the user interactive evaluation process.

[0023] As the preferred embodiment of the present application, in the system screening process, the chromosome individuals that meet the penalty mechanism are set as low scores and considered as invalid individuals, and the rest are considered as valid individuals, and the visual storage layout scheme phenotype is presented to the user for selection and evaluation after decoding the valid individuals.

[0024] As the preferred embodiment of the present application, the crossover rate of the single-point crossover method determines the convergence performance of the population, and too large or too small crossover rate will affect the genotype of the excellent individuals of the population, a mutation operator is introduced, that is, a random mutation operation is performed on a certain gene bit of the parent chromosome to generate a new individual, thereby improving the global search ability of the algorithm.

[0025] Compared with the prior art, the present application has the following advantages: the present application introduces an interactive genetic algorithm for visual scheme design of furniture storage layout in the field of furniture design, and users and designers can break the traditional communication mode to intuitively extract requirements, which not only reduces the cost of manpower and material resources, but also greatly improves the design efficiency and enterprise custom service experience; at the same time, on the basis of the interactive genetic algorithm, a penalty function is introduced, that is, the selected population individuals are punished in the system operation process, which ensures that the storage layout scheme selected and evaluated by the user is reasonable and feasible. A mutation operator is introduced in the algorithm system, which randomly varies a certain gene bit of the parent chromosome to generate a new individual, and when the population iteration is almost convergent and the solution space is too large, the optimal solution of the objective function still does not appear, which improves the global search ability of the algorithm. BRIEF DESCRIPTION OF DRAWINGS

[0026] Figure 1 is the interactive interactive genetic algorithm flowchart of the embodiment of the present application;

[0027] Figure 2 is the customized wardrobe internal storage space grid schematic diagram of the embodiment of the present application;

[0028] Figure 3 is the chromosome coding schematic diagram of the embodiment of the present application;

[0029] Figure 4 is the function module translation rule schematic diagram of the embodiment of the present application;

[0030] Figure 5 is the parent chromosome single-point crossover operation schematic diagram of the embodiment of the present application;

[0031] Figure 6 is the parent chromosome mutation operation schematic diagram of the embodiment of the present application;

[0032] Figure 7 is the best fitness value columnar schematic diagram of the best storage layout scheme of the embodiment of the present application;

[0033] Figure 8 is the iteration number and the best fitness value trend schematic diagram of the best scheme of the embodiment of the present application. DETAILED DESCRIPTION

[0034] The present application is further described below in conjunction with the drawings:

[0035] A furniture individualization customization system and method based on an interactive genetic algorithm, as shown in the figure, comprises the following steps: Figures 1-8

[0036] 1) According to the optimal standard, the specified furniture is divided into function modules.

[0037] 2) The region positioning of different function modules is performed, and the size range is set;

[0038] 3) The grid method is used to build a mathematical model for the function module region, and the letter symbol method is used to sequentially label a~y in the vertical axis direction from left to right;

[0039] 4) The unit module, growth direction and hierarchical variable of the furniture module itself are set as a gene position of the furniture chromosome, which controls the phenotype of each function region of the furniture itself;

[0040] ​5) Following the gene decoding rules, all the gene information contained in the chromosome is decoded in order from a to y in ascending order, until all the gene information is decoded, and the entire furniture storage space layout is covered by the dominant phenotype;

[0041] 6) Set four different functional area limit rules, increase the penalty term of the objective function;

[0042] 7) The initial population size is selected as M=60P55, the initial population P54 is obtained, the fitness is evaluated, and M=60 is selected as the initial population size, and the initial population is evaluated.

[0043] 8) Single-point crossover method is used for gene crossover and mutation operation, and the population crossover rate and mutation rate are set as Pc=0.8, Pm=0.1;

[0044] 9) Output the optimal scheme or the visual furniture scheme when the user evaluation times reach forty-five times.

[0045] A1 standard is the optimal standard, which divides the furniture into multiple functional areas, and sets the storage content and commonly used accessories for each functional area.

[0046] The interior space of the furniture is divided into five areas (A-E): A area is to set the space height in the range of 600-1400mm; B area is to set the space height in the range of 1400-1800mm; C area is to set the space height in the range of above 1800mm; D and E areas are to set the space height in the range of below 600mm.

[0047] According to different functional modules, corresponding space size range is set to meet the use scene of the items in the area in actual life.

[0048] The cells are numbered, and the code is n, where n cell n=i[a≤i≤y] is the initial growth point of the functional module phenotype gene information, which is represented by a three-element array (Ai, Bi, Ci), where A represents the single modulus variable, B represents the growth direction variable, and C represents the level variable. The length of the complete chromosome genotype is 25*3=75 bits, and the acquisition method is to sequentially connect the gene bit information in the a-y cell. In the chromosome individual gene, the phenotype of eight functional modules is represented in the form of a three-element array, and the functional module translation rule is made into a table.

[0049] The phenotype is divided into two types, namely, the recessive phenotype and the dominant phenotype. The recessive phenotype is hidden and invisible during the user's participation in the interactive genetic algorithm, while the dominant phenotype is visible. The decoding priority is used as the basis for the recessive or dominant phenotype of the eight modules. In the a-y cells, the earlier the cell letter number is in the English letter sequence table, the more priority the decoding order has. Here, only the dominant phenotype of the overall furniture storage layout scheme is presented to the user for scoring in a visual effect.

[0050] A penalty function is introduced, and a penalty mechanism is developed. Except for the selected dominant phenotype, the fitness of the remaining same dominant phenotype chromosome individuals is set to 0 and hidden, avoiding participation in the user interaction evaluation process.

[0051] During the system screening process, the chromosome individuals that meet the penalty mechanism are set to a low score and are considered invalid individuals. The remaining individuals are considered valid individuals. After decoding the valid individuals, the visual storage layout scheme phenotype is presented to the user for selection and evaluation.

[0052] The crossover rate of the single-point crossover method determines the convergence performance of the population. Too large or too small crossover rates can affect the genotype of the excellent individuals in the population. A mutation operator is introduced, which randomly mutates a certain gene position of the parent chromosome to generate new individuals, thereby improving the global search ability of the algorithm.

[0053] In use, taking a wardrobe as an example, the storage layout requirements of the custom wardrobe are quite extensive for the general population. Compared with the general population, the elderly and children have special characteristics, and there are also special rule restrictions in the design process of the custom wardrobe. Therefore, the general population is selected as the main test group, and the age and gender distribution is uniform, i.e., the users aged 20 to 55 years old are set as the main population of the custom wardrobe test, and the number of test personnel is 6 and is grouped and labeled. The specific information is as follows:

[0054] Age range Gender Number 20-29 years old Male No. 1 20-29 years old Female No. 2 30-39 years old Male No. 3 30-39 years old Female No. 4 40-55 years old Male No. 5 40-55 years old Female No. 6

[0055] The test group uses the Likert five-level scale as the scoring standard: 0-20 points represent very dissatisfied, 20-40 points represent relatively dissatisfied, 40-60 points represent generally satisfied, 60-80 points represent relatively satisfied, and 80-100 points represent very satisfied. When the user scores the storage layout scheme more than 95 points, the system automatically determines that the scheme is the optimal scheme, and the algorithm is terminated and the optimal scheme is output. When the user participates in the selection and evaluation for 45 times, the system will select the current optimal scheme according to the user evaluation result, and the algorithm will also be automatically terminated.

[0056] In the algorithm operating system, the elite selection method is used for population interaction genetic evolution: in the user participation selection evaluation process, when the score is lower than 20, the individual will be automatically excluded, when the score is between 20-80, the system will introduce the mutation operator for cross variation operation, when the score is higher than 80, the individual will be reserved.

[0057] The selected 6 subjects participate in the evaluation frequency of 40 to 45 times, wherein the highest fitness value is 95, and the average fitness value is 89.7 (as shown in Figure 7 The data analysis shows that the interactive genetic algorithm is reasonable and feasible in the field of furniture design for the layout design of furniture internal storage, and the user can intuitively and efficiently obtain a satisfactory scheme in a small number of selection and evaluation processes.

[0058] The data of the selected subject No. 1 is analyzed, and in the selection and evaluation scheme process, the best fitness value of each generation individual has certain fluctuation, but as the evaluation frequency is more and more, the evaluation fitness value is more and more stable, which also proves that when the user selects and evaluates more and more, the scheme output by the system through the algorithm is more and more in line with the user's demand.

[0059] The beneficial effects of the present application are as follows: the present application introduces an interactive genetic algorithm for visual scheme design in the field of furniture design for furniture storage layout, users and designers can break the traditional communication mode and extract intuitive requirements, which not only reduces the cost of manpower and material resources, but also greatly improves the design efficiency and enterprise custom service experience; at the same time, on the basis of the interactive genetic algorithm, a penalty function is introduced, that is, the selected population individuals are punished in the system operation process, so that the user participation selection and evaluation storage layout scheme has rationality and feasibility. In the algorithm system, a mutation operator is introduced, a certain gene site of the parent chromosome is randomly mutated, a new individual is generated, when the population iteration is nearly convergent and the solution space is too large, the optimal solution of the objective function still does not appear, the global search ability of the algorithm is improved.

[0060] The basic principles, main features and advantages of the present application are shown and described. Those skilled in the art should understand that the present application is not limited to the above examples, the above examples and descriptions in the specification are only to illustrate the principles of the present application, without departing from the spirit and scope of the present application, the present application can have various changes and improvements, and these changes and improvements all fall within the scope of the claimed present application.

Claims

1. A method for personalized furniture customization based on an interactive genetic algorithm, characterized in that, Includes the following steps: According to the A1 standard, the designated furniture functional modules are divided; Different functional modules are located in specific areas, and their size ranges are defined. A mathematical model for the functional module area is built using the grid method, and the letter symbol method is used to label it sequentially from left to right along the vertical axis using a to y. The unit module, growth direction, and hierarchical variables are set for the furniture module itself, which are respectively used as a gene locus of the furniture chromosome to control the phenotype of each functional area of ​​the furniture itself; Following the rules of gene decoding, all the gene information contained in the chromosome is decoded sequentially in ascending order from a to y to the internal storage space cells of the furniture until all gene information is decoded and the entire furniture storage space layout is filled with dominant phenotypes. Four different functional area restriction rules are set, and a penalty term is added to the objective function; We selected M=60 as the initial population size and evaluated the fitness of the initial population. Gene crossover and mutation were performed using the single-point crossover method, with the population crossover rate and mutation rate set to Pc=0.8 and Pm=0.1, respectively. When a user scores a storage layout scheme with a score of over 95, the system automatically determines that the scheme is the optimal one, terminates the algorithm, and outputs the optimal scheme. When the number of times a user participates in the selection and evaluation reaches forty-five, the system will filter out the current optimal scheme based on the user's evaluation results, and the algorithm will also automatically terminate.

2. The method for personalized furniture customization based on an interactive genetic algorithm according to claim 1, characterized in that, The A1 standard is the optimal standard, which divides furniture into multiple functional areas and sets the storage contents and commonly used accessories for each functional area accordingly.

3. The method for personalized furniture customization based on an interactive genetic algorithm according to claim 1, characterized in that, The interior space of the furniture is divided into five zones: Zone A has a height of 600-1400mm; Zone B has a height of 1400-1800mm; Zone C has a height of more than 1800mm; and Zones D and E have a height of less than 600mm.

4. The method for personalized furniture customization based on an interactive genetic algorithm according to claim 3, characterized in that, The corresponding space size range is set according to different functional modules to match the actual usage scenarios of items in that area in real life.

5. The method for personalized furniture customization based on an interactive genetic algorithm according to claim 4, characterized in that, The cells are numbered and coded as n, where cell n = i [a ≤ i ≤ y] serves as the initial growth point. The phenotypic gene information of the functional module is represented by a ternary array (Ai, Bi, Ci), where Ai represents the unit modulus variable, Bi represents the growth direction variable, and Ci represents the hierarchy variable.

6. The method for personalized furniture customization based on interactive genetic algorithm according to claim 1, characterized in that the length of the complete chromosome genotype is 25*3=75 bits, and the method of obtaining it is to sequentially connect the gene position information in cells a to y in order.

7. The method for personalized furniture customization based on an interactive genetic algorithm according to claim 6, characterized in that, The phenotypes of eight functional modules are represented in the form of triplet arrays in the genes of an individual chromosome, and the translation rules of the functional modules are tabulated.

8. The method for personalized furniture customization based on an interactive genetic algorithm according to claim 7, characterized in that, Phenotyps are divided into two types: recessive and manifest. Recessive phenotypes are hidden and invisible during user interaction with the genetic algorithm, while dominant phenotypes are visible. Decoding priority is used as the hidden priority of the eight modules. The basis for determining the dominant or dominant phenotype is that, within cells a to y, the earlier the cell letter appears in the alphabetical order, the higher the decoding priority. Here, only the overall furniture storage layout scheme generated by the dominant phenotype is presented to the user for scoring in a visual manner.

9. A method for personalized furniture customization based on an interactive genetic algorithm according to claim 8, characterized in that, A penalty function is introduced and a penalty mechanism is established. Except for the selected dominant phenotype, the system sets the adaptation of all other individuals with the same dominant phenotype chromosome to 0 and hides them to avoid their participation in the user interaction evaluation process.

10. A method for personalized furniture customization based on an interactive genetic algorithm according to claim 9, characterized in that, During the system screening process, individuals with chromosomes that match the penalty mechanism are set as low-scoring and considered invalid individuals, while the rest are considered valid individuals. After decoding the valid individuals, the visual storage layout scheme is presented to the user for selection and evaluation.

11. The method for personalized furniture customization based on an interactive genetic algorithm according to claim 1, characterized in that, The crossover rate of the single-point crossover method determines the convergence performance of the population. However, if the crossover rate is too high or too low, it will affect the genotype of the superior individuals in the population. The introduction of the mutation operator, that is, the random mutation operation on a certain gene position of the parent chromosome to generate new individuals, thereby improving the global search capability of the algorithm.

Images