An automatic generation method for core tube plan layout of high-rise buildings based on NSGA-Ⅲ

Abstract

A method for automatically generating the plane layout of the core tube of a high-rise building based on NSGA-Ⅲ comprises the following steps: 1) constructing a multi-objective model; 2) encoding the position coordinates and orientation of each component in the core tube, and generating an antechamber corresponding to each component and / or an antechamber shared by multiple components according to the encoding; 3) solving the position coordinates and orientation of each component using the NSGA-Ⅲ algorithm to obtain a non-dominant solution set; 4) automatically generating a core tube plane layout diagram based on the multi-objective model according to the non-dominant solution set and the antechamber obtained in step 2). The present invention proposes a method for automatically generating the plane layout of the core tube of a high-rise building based on NSGA-Ⅲ, fully investigates relevant design specifications and design requirements, formulates reasonable objective functions and constraints, and adopts a non-dominated genetic algorithm based on reference points to quickly obtain a mutually non-dominated optimal solution set that meets the constraints, which can meet the design specifications while also adapting to the diversity and flexibility requirements in different scenarios.

Description

Technical Field

[0001] The present invention relates to the field of information technology, and in particular to a method for automatically generating a core tube plane layout of a high-rise building based on NSGA-III. Background Art

[0002] The core tube is usually located in the center of the building. It consists of stairs, ordinary elevators, fire elevators, antechambers, elevator halls, corridors, pipe shafts, etc. It is the horizontal and vertical space for personnel transportation and equipment layout in the building, and plays an important role in high-rise buildings. Therefore, determining a reasonable and effective core tube is one of the important links in the design of high-rise buildings. However, the traditional core tube plane layout design relies heavily on the designer's experience. Because it is done manually, the whole process is inefficient and prone to errors.

[0003] In recent years, many methods for automatically generating floorplans based on intelligent algorithms have emerged. These methods can avoid human errors and greatly improve design efficiency. Many floorplan design problems belong to NP (uncertain polynomial) problems, and it is difficult to find high-quality solutions. At the same time, it is noted that the core tube floorplan automatic generation problem is essentially a design problem, and there are often more than one goal to be optimized. Moreover, methods based on single-objective optimization will only produce one result and cannot meet the diversity requirements. Summary of the invention

[0004] The purpose of the present invention is to provide a method for automatically generating the plane layout of the core tube of a high-rise building based on NSGA-III, comprising the following steps:

[0005] 1) Construct a multi-objective model.

[0006] 2) Encode the position coordinates and orientation of each component in the core tube, and generate an antechamber corresponding to each component and / or an antechamber shared by multiple components based on the encoding.

[0007] 3) Use the NSGA-Ⅲ algorithm to solve the position coordinates and orientation of each component and obtain the uncontrolled solution set.

[0008] 4) According to the uncontrolled solution set obtained in step 3) and the antechamber obtained in step 2), a core tube floor plan is automatically generated based on a multi-objective model.

[0009] Furthermore, the objective functions f1 and f2 of the multi-objective model are as follows:

[0010] minf1=S traffic_core +S minrec (1)

[0011] maxf2=N shared_room (2)

[0012] In the formula, S traffic_core represents the actual area of ​​the core tube, S min_rec Indicates the minimum unit area of ​​the core tube containing all components, N shared_room Indicates the number of shared antechambers.

[0013] Furthermore, the constraints g1 and g2 of the multi-objective model are as follows:

[0014] stg1=S overlap_items +S averlap_fr +S overlap_cor =0 (3)

[0015] g2=S fire_constrains =0 (4)

[0016] In the formula, S overlap_items represents the overlapping area between components, S overlap_fr represents the overlapping area between each antechamber and the components, S overlap_cor represents the overlapping area of ​​the corridor and each component, S fire_constrains Represents a judgment function, which contains multiple judgment conditions.

[0017] The judgment condition includes whether the area size of the antechamber of each component and the shared antechamber meets the fire protection regulations.

[0018] Furthermore, the step of encoding the position coordinates and orientation of each component in the core tube and generating an antechamber corresponding to each component and / or an antechamber shared by multiple components according to the encoding includes:

[0019] 2.1) Construct a rectangular coordinate system and set the location coordinates and orientation of the stairs to fixed values.

[0020] In the rectangular coordinate system, the coordinates of the lower left corner of the ith component are defined as (x i ,y i ), direction is dir i , i is a positive integer.

[0021] 2.2) Simulate the relative position and orientation of each component in the core tube. The simulation variables are as follows:

[0022] {x1,y1,dir1,x2,y2,dir2,…,x n ,y n ,dir n}(5)In the formula, n is the total number of components.

[0023] 2.3) Set the upper and lower limits of the simulation variables and generate the antechamber of each component according to the current position and orientation.

[0024] If multiple antechambers overlap or are adjacent to each other, the smallest unit containing these antechambers shall be taken as the shared antechamber.

[0025] Furthermore, the step of using the NSGA-III algorithm to solve the position coordinates and orientation of each component to obtain an uncontrolled solution set includes:

[0026] 3.1) Initialize the population and the number of iterations q, and set 2p parent individuals X (x1, x2, ..., x n1 ), n1 is the number of parameters contained in the variable, p and q are both positive integers; the initial value of the iteration number q is 1.

[0027] 3.2) Make p parent individuals cross over with the remaining p parent individuals in pairs to produce 2p offspring individuals; the crossover method includes the SBX method.

[0028] Two parent individuals cross to produce two corresponding offspring individuals and As shown below:

[0029]

[0030]

[0031] Where i1 is the parameter number in the individual and n1 is the number of parameters contained in the individual.

[0032] Offspring and Parameters in x i1 ,y i1 are the parent individuals X(x1,x2,…,x n1 )、Y(y1,y2,…,y n1 ) in the .

[0033] Among them, the intermediate parameter β is as follows:

[0034]

[0035] Where η is the distribution factor. rand is a random number, rand∈[0,1).

[0036] 3.3) The polynomial mutation method is used to mutate the offspring individuals to obtain the offspring individuals, as shown below:

[0037]

[0038] In the formula, is the i1th parameter of the mutated offspring individual. is the i1th parameter of the offspring individual before mutation.

[0039] Among them, the intermediate parameter β i1 As shown below:

[0040]

[0041] In the formula, rand i1 is a random number, rand i1 ∈[0,1).

[0042] 3.4) After merging the parent individuals and the offspring individuals, they are sorted according to the dominance relationship and divided into solution sets of different levels.

[0043] 3.5) Let the number of iterations q = q + 1, select 2p individuals from the solution sets of different levels as the new parent population, and return to step 3.2) until the number of iterations q reaches the preset threshold Q, and the undominated solution set is obtained, where Q is a positive integer.

[0044] Furthermore, the step of merging the parent individuals and the child individuals and sorting them according to the dominance relationship includes:

[0045] 3.4.1) Calculate the objective function value f1, f2 and constraint violation value g1, g2 for each individual.

[0046] 3.4.2) Compare individuals pairwise and sort them according to dominance relationship.

[0047] Furthermore, the conditions under which individual X dominates individual Y are as follows:

[0048] ⅰ. If X and Y are both feasible solutions, the objective function value f1(X)>f1(Y) and the objective function value f2(X)>f2(Y).

[0049] ⅱ.X is a feasible solution, Y is an infeasible solution.

[0050] ⅲ. If both X and Y are infeasible solutions, CV(X)>CV(Y), where CV(X)=g1(X)+g2(X), CV(Y)=g1(Y)+g2(Y).

[0051] Among them, a feasible solution refers to a solution that satisfies the constraints, and an infeasible solution refers to a solution that does not satisfy the constraints.

[0052] Furthermore, the step of selecting 2p individuals from the solution sets of different levels as the new parent population includes:

[0053] 3.5.1) Define the number of new parent populations to be 0, and the solution set of the current level to be the solution set that dominates the highest level.

[0054] 3.5.2) Compare whether the number of individuals k in the new parent population is greater than the preset number of individuals 2p. If so, proceed to step 3.5.3), if not, proceed to step 3.5.4).

[0055] 3.5.3) Select individuals from the solution set of the current level according to the reference point-based selection method defined by the NSGA-III algorithm, and select the individuals selected according to the reference point-based selection method defined by the NSGA-III algorithm into the new parent population.

[0056] 3.5.4) Select all individuals in the solution set of the current level into the new parent population, and make the solution set of the next level as the solution set of the current level, and return to step 3.5.2).

[0057] Further, the step of selecting individuals from the solution set of the current level according to the reference point-based selection method defined by the NSGA-III algorithm, and selecting the individuals selected according to the reference point-based selection method defined by the NSGA-III algorithm into the new parent population includes:

[0058] 3.5.3.1) Standardizing the solution set of the current level, including the following steps:

[0059] 3.5.3.1.1) Find the ideal point of the solution set of the current level, which is expressed as in, It is the minimum value of the jth objective function of all individuals in the solution set of the current level, j∈[1,m], and m is the number of objective functions.

[0060] 3.5.3.1.2) Translate the objective function of each individual in the solution set of the current level so that the ideal point becomes the origin and construct a coordinate system.

[0061] The j-th objective function of each individual in the solution set of the current level is as follows:

[0062]

[0063] In the formula, f ′ j (x) is the objective function after translation. j (x) is the objective function before translation.

[0064] 3.5.3.1.3) Calculate the extreme points of each objective function in the solution set of the current level as follows:

[0065]

[0066] In the formula, is the extreme point of the jth objective function in the solution set of the current level. w is the unit vector of each objective function.j is the unit vector of the jth objective function in the solution set of the current level.

[0067] 3.5.3.1.4) Construct a hyperplane based on the extreme points and set reference points evenly on the hyperplane Calculate the intercept a of the hyperplane with each coordinate axis j .

[0068] 3.5.3.1.5) Normalize the j-th objective function in the solution set of the current level as follows:

[0069]

[0070] In the formula, f j n (x) is the objective function after normalization.

[0071] 3.5.3.2) Associating the standardized individuals to the corresponding reference points, including the following steps:

[0072] 3.5.3.2.1) Each reference point and the corresponding origin Connect the lines to get the reference line W j .

[0073] 3.5.3.2.2) Traverse each standardized individual s, find the reference line W that is closest to individual s, and find the reference point π(s) that is closest to individual s on the reference line W, and calculate the distance d between individual s and the reference point π(s) ⊥ (s,π(s)), as follows:

[0074]

[0075] 3.5.3.3) Select the reference point π(s) with the least number of associated individuals. If there are multiple reference points π(s) with the least number of associated individuals, randomly select a reference point S from them and set the set of all individuals associated with the reference point S as I S .

[0076] 3.5.3.4) Judgment set I S Is it empty? If so, return to step 3.5.3.3); if not, go to step 3.5.3.5).

[0077] 3.5.3.5) Set I S The number of individuals of all levels before the solution set of the current level is ρ S , determine the number of individuals ρ S = 0 is true, if so, then from the set I SThe individual with the smallest distance from the reference point S is selected to enter the new parent population. If not, the individual is selected from the set I S An individual is randomly selected from the population to enter the new parent population.

[0078] 3.5.3.6) Determine whether the number of individuals in the new parent population is equal to the preset number of individuals 2p. If so, construct a new parent population. If not, return to step 3.5.3.3).

[0079] Furthermore, the step of automatically generating a core tube plan layout diagram based on a multi-objective model according to the uncontrolled solution set obtained in step 3) and the front room obtained in step 2) includes:

[0080] 4.1) Determine whether the boundary of the antechamber meets the conditions for opening doors and windows. If so, add doors and windows at the corresponding boundary of the antechamber, and generate a wind shaft at the boundary of the corridor or inside the antechamber.

[0081] 4.2) Generate the core tube floor plan based on the constraints of the multi-objective model.

[0082] The technical effect of the present invention is unquestionable. The present invention proposes an automatic generation method for the core tube plane layout of a high-rise building based on NSGA-Ⅲ, fully investigates relevant design specifications and design requirements, formulates reasonable objective functions and constraints, and adopts a non-dominated genetic algorithm based on reference points to quickly obtain a mutually non-dominated optimal solution set that meets the constraints. While meeting the design specifications, it can also adapt to the diversity and flexibility requirements in different scenarios.

[0083] The present invention conducts an in-depth investigation on the design requirements and specifications of the core tube of high-rise residential buildings, abstracts the core tube component attributes and component relationships into multiple objectives and multiple constraints, and combines the NSGA-Ⅲ intelligent optimization algorithm to meet the fire protection and general building design specifications while having high design diversity and flexibility to meet the design requirements in different scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0084] Figure 1 The present invention discloses a method for automatically generating the plane layout of the core tube of a high-rise building based on NSGA-Ⅲ;

[0085] Figure 2 It is a schematic diagram of the position and orientation coding of each component in the core tube of the present invention;

[0086] Figure 3 The present invention generates an antechamber and a schematic diagram of a use rule according to the coding;

[0087] Figure 4 This is the flow chart of the NSGA-III algorithm of the present invention;

[0088] Figure 5Result diagram of the core tube of the high-rise building designed for this invention. DETAILED DESCRIPTION

[0089] The present invention is further described below in conjunction with the embodiments, but it should not be understood that the above subject matter of the present invention is limited to the following embodiments. Without departing from the above technical ideas of the present invention, various substitutions and changes are made according to the common technical knowledge and customary means in the art, which should all be included in the protection scope of the present invention.

[0090] Embodiment 1:

[0091] See also Figures 1 to 5 , a method for automatically generating the core tube plane layout of a high-rise building based on NSGA-Ⅲ, comprising the following steps:

[0092] 1) Construct a multi-objective model.

[0093] 2) Encode the position coordinates and orientation of each component in the core tube, and generate an antechamber corresponding to each component and / or an antechamber shared by multiple components based on the encoding.

[0094] 3) Use the NSGA-Ⅲ algorithm to solve the position coordinates and orientation of each component and obtain the uncontrolled solution set.

[0095] 4) According to the uncontrolled solution set obtained in step 3) and the antechamber obtained in step 2), a core tube floor plan is automatically generated based on a multi-objective model.

[0096] Embodiment 2:

[0097] A method for automatically generating the core tube plane layout of a high-rise building based on NSGA-III, the main technical content of which is shown in Example 1. Furthermore, the objective functions f1 and f2 of the multi-objective model are as follows:

[0098]

[0099] In the formula, S traffic_core represents the actual area of ​​the core tube, S min_rec Indicates the minimum unit area of ​​the core tube containing all components, N shared_room Indicates the number of shared antechambers.

[0100] Embodiment 3:

[0101] A method for automatically generating a core tube plane layout of a high-rise building based on NSGA-III, the main technical content of which is shown in any one of Embodiments 1 to 2. Further, the constraints g1 and g2 of the multi-objective model are as follows:

[0102] stg1=S overlap_items +S overlap_fr +Soverlap_cor =0 (3)

[0103] g2=S fire_constrains =0 (4)

[0104] In the formula, S overlap_items represents the overlapping area between components, S overlap_fr represents the overlapping area between each antechamber and the components, S overlap_cor represents the overlapping area of ​​the corridor and each component, S fire_constrains Represents a judgment function, which contains multiple judgment conditions.

[0105] The judgment condition includes whether the area size of the antechamber of each component and the shared antechamber meets the fire protection regulations.

[0106] Embodiment 4:

[0107] A method for automatically generating a plan layout of a core tube of a high-rise building based on NSGA-III, the main technical content of which is shown in any one of embodiments 1 to 3. Further, the step of encoding the position coordinates and orientation of each component in the core tube and generating an antechamber corresponding to each component and / or an antechamber shared by multiple components according to the encoding comprises:

[0108] 2.1) Construct a rectangular coordinate system and set the location coordinates and orientation of the stairs to fixed values.

[0109] In the rectangular coordinate system, the coordinates of the lower left corner of the ith component are defined as (x i ,y i ), direction is dir i , i is a positive integer.

[0110] 2.2) Simulate the relative position and orientation of each component in the core tube. The simulation variables are as follows:

[0111] {x1,y1,dir1,x2,y2,dir2,…,x n ,y n ,dir n} (5)

[0112] Where n is the total number of components.

[0113] 2.3) Set the upper and lower limits of the simulation variables and generate the antechamber of each component according to the current position and orientation.

[0114] If multiple antechambers overlap or are adjacent to each other, the smallest unit containing these antechambers shall be taken as the shared antechamber.

[0115] Embodiment 5:

[0116] A method for automatically generating a plan layout of a core tube of a high-rise building based on NSGA-III, the main technical content of which is shown in any one of embodiments 1 to 4. Further, the step of using the NSGA-III algorithm to solve the position coordinates and orientation of each component to obtain an uncontrolled solution set includes:

[0117] 3.1) Initialize the population and the number of iterations q, and set 2p parent individuals X (x1, x2, ..., x n1 ), n1 is the number of parameters contained in the variable, p and q are both positive integers; the initial value of the iteration number q is 1.

[0118] 3.2) Make p parent individuals cross over with the remaining p parent individuals in pairs to produce 2p offspring individuals; the crossover method includes the SBX method.

[0119] Two parent individuals cross to produce two corresponding offspring individuals and As shown below:

[0120]

[0121]

[0122] Where i1 is the parameter number in the individual and n1 is the number of parameters contained in the individual.

[0123] Offspring and Parameters in x i1 ,y i1 are the parent individuals X(x1,x2,…,x n1 )、Y(y1,y2,…,y n1 ) in the .

[0124] Among them, the intermediate parameter β is as follows:

[0125]

[0126] Where η is the distribution factor. rand is a random number, rand∈[0,1).

[0127] 3.3) The polynomial mutation method is used to mutate the offspring individuals to obtain the offspring individuals, as shown below:

[0128]

[0129] In the formula, is the i1th parameter of the mutated offspring individual. is the i1th parameter of the offspring individual before mutation.

[0130] Among them, the intermediate parameter β i1 As shown below:

[0131]

[0132] In the formula, rand i1 is a random number, rand i1 ∈[0,1).

[0133] 3.4) After merging the parent individuals and the offspring individuals, they are sorted according to the dominance relationship and divided into solution sets of different levels.

[0134] 3.5) Let the number of iterations q = q + 1, select 2p individuals from the solution sets of different levels as the new parent population, and return to step 3.2) until the number of iterations q reaches the preset threshold Q, and the undominated solution set is obtained, where Q is a positive integer.

[0135] Embodiment 6:

[0136] A method for automatically generating a core tube plane layout of a high-rise building based on NSGA-III, the main technical content of which is shown in any one of Embodiments 1 to 5. Further, the step of merging the parent individual and the child individual and sorting them according to the dominance relationship comprises:

[0137] 3.4.1) Calculate the objective function value f1, f2 and constraint violation value g1, g2 for each individual.

[0138] 3.4.2) Compare individuals pairwise and sort them according to dominance relationship.

[0139] Embodiment 7:

[0140] A method for automatically generating a core tube plane layout of a high-rise building based on NSGA-III, the main technical content of which is shown in any one of Examples 1 to 6. Furthermore, the condition that individual X dominates individual Y is as follows:

[0141] ⅰ. If X and Y are both feasible solutions, the objective function value f1(X)>f1(Y) and the objective function value f2(X)>f2(Y).

[0142] ⅱ.X is a feasible solution, Y is an infeasible solution.

[0143] ⅲ. If both X and Y are infeasible solutions, CV(X)>CV(Y), where CV(X)=g1(X)+g2(X), CV(Y)=g1(Y)+g2(Y).

[0144] Among them, a feasible solution refers to a solution that satisfies the constraints, and an infeasible solution refers to a solution that does not satisfy the constraints.

[0145] Embodiment 8:

[0146] A method for automatically generating a plan layout of a core tube of a high-rise building based on NSGA-III, the main technical content of which is shown in any one of Examples 1 to 7. Further, the step of selecting 2p individuals from solution sets of different levels as a new parent population comprises:

[0147] 3.5.1) Define the number of new parent populations to be 0, and the solution set of the current level to be the solution set that dominates the highest level.

[0148] 3.5.2) Compare whether the number of individuals k in the new parent population is greater than the preset number of individuals 2p. If so, proceed to step 3.5.3), if not, proceed to step 3.5.4).

[0149] 3.5.3) Select individuals from the solution set of the current level according to the reference point-based selection method defined by the NSGA-III algorithm, and select the individuals selected according to the reference point-based selection method defined by the NSGA-III algorithm into the new parent population.

[0150] 3.5.4) Select all individuals in the solution set of the current level into the new parent population, and make the solution set of the next level as the solution set of the current level, and return to step 3.5.2).

[0151] Embodiment 9:

[0152] A method for automatically generating a plan layout of a core tube of a high-rise building based on NSGA-III, the main technical content of which is shown in any one of embodiments 1 to 8. Further, the step of selecting an individual from a solution set of a current level according to a reference point-based selection method defined by the NSGA-III algorithm, and selecting the individual selected by the reference point-based selection method defined by the NSGA-III algorithm into a new parent population comprises:

[0153] 3.5.3.1) Standardizing the solution set of the current level, including the following steps:

[0154] 3.5.3.1.1) Find the ideal point of the solution set of the current level, which is expressed as in, It is the minimum value of the jth objective function of all individuals in the solution set of the current level, j∈[1,m], and m is the number of objective functions.

[0155] 3.5.3.1.2) Translate the objective function of each individual in the solution set of the current level so that the ideal point becomes the origin and construct a coordinate system.

[0156] The j-th objective function of each individual in the solution set of the current level is as follows:

[0157]

[0158] Where f′ j (x) is the objective function after translation. j (x) is the objective function before translation.

[0159] 3.5.3.1.3) Calculate the extreme points of each objective function in the solution set of the current level as follows:

[0160]

[0161] In the formula, is the extreme point of the jth objective function in the solution set of the current level. w is the unit vector of each objective function. j is the unit vector of the jth objective function in the solution set of the current level.

[0162] 3.5.3.1.4) Construct a hyperplane based on the extreme points and set reference points evenly on the hyperplane Calculate the intercept a of the hyperplane with each coordinate axis j .

[0163] 3.5.3.1.5) Normalize the j-th objective function in the solution set of the current level as follows:

[0164]

[0165] In the formula, f j n (x) is the objective function after normalization.

[0166] 3.5.3.2) Associating the standardized individuals to the corresponding reference points, including the following steps:

[0167] 3.5.3.2.1) Each reference point and the corresponding origin Connect the lines to get the reference line W j .

[0168] 3.5.3.2.2) Traverse each standardized individual s, find the reference line W that is closest to individual s, and find the reference point π(s) that is closest to individual s on the reference line W, and calculate the distance d between individual s and the reference point π(s) ⊥ (s,π(s)), as follows:

[0169] d ⊥ (s,π(s))=||(sW T sW / ||W|| 2 )|| (14)

[0170] 3.5.3.3) Select the reference point π(s) with the least number of associated individuals. If there are multiple reference points π(s) with the least number of associated individuals, randomly select a reference point S from them and set the set of all individuals associated with the reference point S as I S .

[0171] 3.5.3.4) Judgment set I S Is it empty? If so, return to step 3.5.3.3); if not, go to step 3.5.3.5).

[0172] 3.5.3.5) Set I S The number of individuals of all levels before the solution set of the current level is ρ S , determine the number of individuals ρ S = 0 is true, if so, then from the set I S The individual with the smallest distance from the reference point S is selected to enter the new parent population. If not, the individual is selected from the set I S An individual is randomly selected from the population to enter the new parent population.

[0173] 3.5.3.6) Determine whether the number of individuals in the new parent population is equal to the preset number of individuals 2p. If so, construct a new parent population. If not, return to step 3.5.3.3).

[0174] Embodiment 10:

[0175] A method for automatically generating a core tube plane layout of a high-rise building based on NSGA-III, the main technical content of which is shown in any one of embodiments 1 to 9. Further, the step of automatically generating a core tube plane layout diagram based on a multi-objective model according to the non-controllable solution set obtained in step 3) and the antechamber obtained in step 2) comprises:

[0176] 4.1) Determine whether the boundary of the antechamber meets the conditions for opening doors and windows. If so, add doors and windows at the corresponding boundary of the antechamber, and generate a wind shaft at the boundary of the corridor or inside the antechamber.

[0177] 4.2) Generate the core tube floor plan based on the constraints of the multi-objective model.

[0178] Embodiment 11:

[0179] See also Figures 1 to 5 , a method for automatically generating the core tube plane layout of a high-rise building based on NSGA-Ⅲ, comprising the following steps:

[0180] 1) Construct a multi-objective model.

[0181] According to the core tube design experience and diversity requirements, two objective functions are set to evaluate the pros and cons of the core tube design. According to the fire protection and ordinary residential design requirements, two constraints are set to make the core tube design specifications available.

[0182] 2) Encode the position coordinates and orientation of each component in the core tube, and generate an antechamber corresponding to each component and / or an antechamber shared by multiple components based on the encoding.

[0183] 3) Use the NSGA-Ⅲ algorithm to solve the position coordinates and orientation of each component and obtain the uncontrolled solution set.

[0184] 4) According to the uncontrolled solution set obtained in step 3) and the antechamber obtained in step 2), a core tube floor plan is automatically generated based on a multi-objective model.

[0185] Embodiment 12:

[0186] A method for automatically generating the core tube plane layout of a high-rise building based on NSGA-Ⅲ, the main technical content of which is shown in Example 11. Furthermore, the objective functions are set as the core tube area target (including the actual area of ​​the core tube and the minimum rectangular area that can contain each component) and the target of the number of shared antechambers.

[0187] The objective functions f1 and f2 of the multi-objective model are as follows:

[0188] minf1=S traffic_core +S minrec (1)

[0189] maxf2=N shared_room (2)

[0190] In the formula, S traffic_core represents the actual area of ​​the core tube, S min_rec Indicates the minimum unit area of ​​the core tube containing all components, N shared_room Indicates the number of shared antechambers.

[0191] Embodiment 13:

[0192] A method for automatically generating a core tube plan layout of a high-rise building based on NSGA-III, wherein the main technical contents are shown in any one of Examples 11 to 12, and further, the constraints are respectively core tube availability constraints (including constraints such as overlapping areas of various components) and normative constraints (including fire protection constraints such as safe exit distances and the size of the antechamber area of ​​various components).

[0193] The constraints g1 and g2 of the multi-objective model are as follows:

[0194] stg1=S overlap_items +Soverlap_fr +S overlap_cor =0 (3)

[0195] g2=S fire_constrains =0 (4)

[0196] In the formula, S overlap_items represents the overlapping area between components, S overlap_fr represents the overlapping area between each antechamber and the components, S overlap_cor represents the overlapping area of ​​the corridor and each component, S fire_constrains Represents a judgment function, which contains multiple judgment conditions.

[0197] The judgment condition includes whether the area size of the antechamber of each component and the shared antechamber meets the fire protection regulations.

[0198] The calculation method of g1 and g2 is: if the specification is met, 0 is returned; otherwise, the difference between the specification value and the actual value is returned.

[0199] Embodiment 14:

[0200] A method for automatically generating a plan layout of a core tube of a high-rise building based on NSGA-III, the main technical content of which is shown in any one of Examples 11 to 13. Further, the step of encoding the position coordinates and orientation of each component in the core tube and generating an antechamber corresponding to each component and / or an antechamber shared by multiple components according to the encoding comprises:

[0201] 2.1) Construct a rectangular coordinate system and set the location coordinates and orientation of the stairs to fixed values.

[0202] In the rectangular coordinate system, the coordinates of the lower left corner of the ith component are defined as (x i ,y i ), direction is dir i , i is a positive integer.

[0203] 2.2) Simulate the relative position and orientation of each component in the core tube. The simulation variables are as follows:

[0204] {x1,y1,dir1,x2,y2,dir2,…,x n ,y n ,dir n} (5)

[0205] Where n is the total number of components.

[0206] 2.3) Set the upper and lower limits of the simulation variables and generate the antechamber of each component according to the current position and orientation.

[0207] If multiple antechambers overlap or are adjacent to each other, the smallest unit containing these antechambers shall be taken as the shared antechamber.

[0208] The upper limit of the variable range is set to [0,0,1,0,0,1,0,0,1], and the lower limit is set to [15,15,4,15,15,4,15,15,4]. Figure 3 As shown, after the operator is generated, the antechambers of each component are generated according to the current position and orientation. The rule is that if the antechambers overlap or are adjacent, the smallest rectangle that can contain these antechambers is taken as the shared antechamber.

[0209] Embodiment 15:

[0210] A method for automatically generating a plan layout of a core tube of a high-rise building based on NSGA-III, the main technical content of which is shown in any one of Examples 11 to 14. Further, the step of using the NSGA-III algorithm to solve the position coordinates and orientation of each component to obtain an uncontrolled solution set includes:

[0211] 3.1) Initialize the population and the number of iterations q, and set 1000 parent individuals X (x1, x2, …, x n1 ), n1 is the number of parameters contained in the variable, q is a positive integer. The initial value of the iteration number q is 1.

[0212] Among them, the parent individual X(x1,x2,…,x n1 ) represent the relative positions and directions of the components in the core tube. The parent individual x1 corresponds to the x1 in the simulation variable in step 5), the parent individual x2 corresponds to the y1 in the simulation variable in step 5), the parent individual x3 corresponds to the dir1 in the simulation variable in step 5), the parent individual x4 corresponds to the x2 in the simulation variable in step 5….

[0213] 1000 initial population individuals are randomly generated, and the offspring population size is set to 1000. Then the following iterative process is entered until the number of iterations reaches 2000.

[0214] 3.2) Make 500 parent individuals cross over with the remaining 500 parent individuals to produce 1000 offspring individuals. The crossover method includes the SBX method.

[0215] Two parent individuals cross to produce two corresponding offspring individuals and As shown below:

[0216]

[0217]

[0218] Where i1 is the parameter number in the individual and n1 is the number of parameters contained in the individual.

[0219] Offspring and Parameters in x i1 ,y i1 are the parent individuals X(x1,x2,…,x n1 )、Y(y1,y2,…,y n1 ) in the .

[0220] Among them, the intermediate parameter β is as follows:

[0221]

[0222] Where η is the distribution factor, and the crossover distribution factor η is set to 20. rand is a random number, rand∈[0,1).

[0223] The crossover adopts the SBX method, and the crossover probability is set to 0.9.

[0224] 3.3) The polynomial mutation method is used to mutate the offspring individuals to obtain 1000 offspring individuals, as shown below:

[0225]

[0226] In the formula, is the i1th parameter of the mutated offspring individual. is the i1th parameter of the offspring individual before mutation.

[0227] Among them, the intermediate parameter β i1 As shown below:

[0228]

[0229] In the formula, rand i1 is a random number, rand i1 ∈[0,1).

[0230] The PM method was used for mutation, the mutation probability was set to 0.8, and the distribution factor η of the mutation was set to 15.

[0231] 3.4) After merging the parent individuals and the offspring individuals, they are sorted according to the dominance relationship and divided into solution sets of different levels.

[0232] 3.5) Let the number of iterations q = q + 1, select 1000 individuals from the solution sets of different levels as the new parent population, and return to step 3.2) until the number of iterations q reaches 2000 and the undominated solution set is obtained.

[0233] Embodiment 16:

[0234] A method for automatically generating a core tube plane layout of a high-rise building based on NSGA-III, the main technical content of which is shown in any one of Examples 11 to 15. Further, the step of merging the parent individual and the child individual and sorting them according to the dominance relationship comprises:

[0235] 3.4.1) Calculate the objective function value f1, f2 and constraint violation value g1, g2 for each individual.

[0236] 3.4.2) Compare individuals pairwise and sort them according to dominance relationship.

[0237] The sorting process is to first find the solution set that cannot be dominated by other solutions, divide it into the first layer, and then find the solution set that cannot be dominated by other solutions from the remaining solution set as the second layer, and so on to divide it into different levels. The dominance relationship is determined according to the CDP principle: for each generated individual, calculate the two objective function values ​​f1, f2 and constraint violation values ​​g1, g2, and compare the individuals pairwise.

[0238] Embodiment 17:

[0239] A method for automatically generating a core tube plane layout of a high-rise building based on NSGA-III, the main technical content of which is shown in any one of Examples 11 to 16. Further, the condition that individual X dominates individual Y is as follows:

[0240] ⅰ. If X and Y are both feasible solutions, the objective function value f1(X)>f1(Y) and the objective function value f2(X)>f2(Y).

[0241] ⅱ.X is a feasible solution, Y is an infeasible solution.

[0242] ⅲ. If both X and Y are infeasible solutions, CV(X)>CV(Y), where CV(X)=g1(X)+g2(X), CV(Y)=g1(Y)+g2(Y).

[0243] Among them, a feasible solution refers to a solution that satisfies the constraints, and an infeasible solution refers to a solution that does not satisfy the constraints.

[0244] Embodiment 18:

[0245] A method for automatically generating a plan layout of a core tube of a high-rise building based on NSGA-III, the main technical content of which is shown in any one of Examples 11 to 17. Further, the step of selecting 1000 individuals from solution sets of different levels as a new parent population comprises:

[0246] 3.5.1) Define the number of new parent populations to be 0, and the solution set of the current level to be the solution set that dominates the highest level.

[0247] 3.5.2) Compare whether the number of individuals k in the new parent population is greater than the preset number of individuals 1000. If so, proceed to step 3.5.3), if not, proceed to step 3.5.4).

[0248] 3.5.3) Select individuals from the solution set of the current level according to the reference point-based selection method defined by the NSGA-III algorithm, and select the individuals selected according to the reference point-based selection method defined by the NSGA-III algorithm into the new parent population.

[0249] 3.5.4) Select all individuals in the solution set of the current level into the new parent population, and make the solution set of the next level as the solution set of the current level, and return to step 3.5.2).

[0250] Embodiment 19:

[0251] A method for automatically generating a plan layout of a core tube of a high-rise building based on NSGA-III, the main technical content of which is shown in any one of Examples 11 to 18. Further, the step of selecting an individual from a solution set of a current level according to a reference point-based selection method defined by the NSGA-III algorithm, and selecting the individual selected by the reference point-based selection method defined by the NSGA-III algorithm into a new parent population comprises:

[0252] 3.5.3.1) Standardizing the solution set of the current level, including the following steps:

[0253] 3.5.3.1.1) Find the ideal point of the solution set of the current level, which is expressed as in, It is the minimum value of the jth objective function of all individuals in the solution set of the current level, j∈[1,m], and m is the number of objective functions.

[0254] 3.5.3.1.2) Translate the objective function of each individual in the solution set of the current level so that the ideal point becomes the origin and construct a coordinate system.

[0255] The j-th objective function of each individual in the solution set of the current level is as follows:

[0256]

[0257] In the formula, f′ j (x) is the objective function after translation. j (x) is the objective function before translation.

[0258] 3.5.3.1.3) Calculate the extreme points of each objective function in the solution set of the current level as follows:

[0259]

[0260] In the formula, is the extreme point of the jth objective function in the solution set of the current level. w is the unit vector of each objective function. j is the unit vector of the jth objective function in the solution set of the current level.

[0261] 3.5.3.1.4) Construct a hyperplane based on the extreme points and set reference points evenly on the hyperplane Calculate the intercept a of the hyperplane with each coordinate axis j .

[0262] 3.5.3.1.5) Normalize the j-th objective function in the solution set of the current level as follows:

[0263]

[0264] In the formula, f j n (x) is the objective function after normalization.

[0265] 3.5.3.2) Associating the standardized individuals to the corresponding reference points, including the following steps:

[0266] 3.5.3.2.1) Each reference point and the corresponding origin Connect the lines to get the reference line W j .

[0267] 3.5.3.2.2) Traverse each standardized individual s, find the reference line W that is closest to individual s, and find the reference point π(s) that is closest to individual s on the reference line W, and calculate the distance d between individual s and the reference point π(s) ⊥ (s,π(s)), as follows:

[0268] d ⊥ (s,π(s))=||(sW T sW / ||W|| 2 )|| (14)

[0269] 3.5.3.3) Select the reference point π(s) with the least number of associated individuals. If there are multiple reference points π(s) with the least number of associated individuals, randomly select a reference point S from them and set the set of all individuals associated with the reference point S as I S .

[0270] 3.5.3.4) Judgment set I S Is it empty? If so, return to step 3.5.3.3); if not, go to step 3.5.3.5).

[0271] 3.5.3.5) Set I S The number of individuals of all levels before the solution set of the current level is ρ S , determine the number of individuals ρ S = 0 is true, if so, then from the set I S The individual with the smallest distance from the reference point S is selected to enter the new parent population. If not, the individual is selected from the set I S An individual is randomly selected from the population to enter the new parent population.

[0272] 3.5.3.6) Determine whether the number of individuals in the new parent population is equal to the preset number of individuals 1000. If so, construct a new parent population. If not, return to step 3.5.3.3).

[0273] Embodiment 20:

[0274] A method for automatically generating a core tube plane layout of a high-rise building based on NSGA-III, the main technical content of which is shown in any one of Examples 11 to 19. Further, the step of automatically generating a core tube plane layout diagram based on a multi-objective model according to the uncontrolled solution set obtained in step 3) and the antechamber obtained in step 2) comprises:

[0275] 4.1) After each uncontrolled solution is determined, determine whether the boundary of the antechamber meets the conditions for opening doors and windows. If the conditions are met, add doors and windows to the corresponding antechamber boundary and generate a wind shaft at the corridor boundary or inside the antechamber.

[0276] 4.2) Generate the core tube floor plan based on the constraints of the multi-objective model.

[0277] The position, orientation and size of the antechamber of each component are determined according to the uncontrolled solution.

[0278] Each of the final uncontrolled solutions, together with the air shafts, doors and windows, etc., are checked for constraints and the core tube plan is drawn.

[0279] Embodiment 21:

[0280] See also Figures 1 to 5 , a method for automatically generating the core tube plane layout of a high-rise building based on NSGA-Ⅲ, comprising the following steps:

[0281] S101. Two objective functions are set based on the core tube design experience and diversity requirements to evaluate the pros and cons of the core tube design. Two constraints are set based on the fire protection and ordinary residential design requirements to make the core tube design specifications available.

[0282] S102, encoding the position coordinates and respective orientations of the main components in the core tube, and generating respective antechambers and a shared antechamber according to the encoding.

[0283] S103. Use the NSGA-III algorithm to optimize the constrained multi-objective task defined in S101. After initializing the population, perform iterations. The iterative process includes initializing the population, crossover and mutation, sorting and grading according to the dominance relationship, selecting subpopulations using a reference point-based method, and repeating the above steps until the maximum number of iterations is reached.

[0284] S104. Add each uncontrolled solution finally generated in S103 to the air shaft, doors and windows, etc., and after checking the constraints, draw the core tube plan layout diagram.

[0285] Step S101 includes:

[0286] According to the design requirements and design specifications, the objective function and constraints are set. The objective functions are set as the core tube area target (including the actual area of ​​the core tube and the minimum rectangular area that can contain each component) and the target of the number of shared antechambers. The constraints are the core tube availability constraints (including constraints such as the overlapping area of ​​each component) and the normative constraints (including fire protection constraints such as the distance to the safe exit and the size of the antechamber area of ​​each component).

[0287] Step S102 includes:

[0288] The coordinates of the lower left corner of the stairs, ordinary elevators, fire elevators and pipe shafts, as well as the orientations of ordinary elevators, fire elevators and pipe shafts are encoded to simulate the relative position and orientation of each component in the core tube. Based on the position and orientation coding, the antechambers of each component and the shared antechamber are generated.

[0289] Step S103 includes:

[0290] The NSGA-Ⅲ method is used to optimize the multi-objective task with constraints, including selection, crossover, mutation and other steps. Among them, the crossover adopts the SBX method, the crossover probability is set to 0.9, the mutation adopts the PM method, the mutation probability is set to 0.8, and the selection strategy is the reference point-based method defined in the NSGA-Ⅲ algorithm. The constraint processing method adopts the CDP method with separation of objective function and constraint conditions.

[0291] Step S104 includes:

[0292] After each uncontrolled solution is determined, check the boundary of the antechamber of the core tube or the shared antechamber. After the conditions for door and window placement are met, add doors and windows to the components that need to open doors and windows, then generate air shafts at the boundary of the corridor or inside the antechamber, and finally draw the results.

[0293] Embodiment 22:

[0294] See also Figures 1 to 5 , a method for automatically generating the core tube plane layout of a high-rise building based on NSGA-Ⅲ, comprising the following steps:

[0295] S101. Two objective functions are set based on the core tube design experience and diversity requirements to evaluate the pros and cons of the core tube design. Two constraints are set based on the fire protection and ordinary residential design requirements to make the core tube design specifications available.

[0296] S102, encoding the position coordinates and respective orientations of the main components in the core tube, and generating respective antechambers and a shared antechamber according to the encoding.

[0297] S103. Use the NSGA-III algorithm to optimize the constrained multi-objective task defined in S101. After initializing the population, perform iterations. The iterative process includes initializing the population, crossover and mutation, sorting and grading according to the dominance relationship, selecting subpopulations using a reference point-based method, and repeating the above steps until the maximum number of iterations is reached.

[0298] S104. Add each uncontrolled solution finally generated in S103 to the air shaft, doors and windows, etc., and after checking the constraints, draw the core tube plan layout diagram.

[0299] Step S101 includes:

[0300] According to actual design experience and diverse needs, two objective functions are set, namely the core tube area target (including the actual area of ​​the core tube and the minimum rectangular area that can contain each component) and the target of the number of shared antechambers; according to fire protection regulations and ordinary residential design regulations, two constraints are set, namely the core tube availability constraint (including constraints such as the overlapping area of ​​each component) and normative constraints (including fire protection constraints such as the distance to the safe exit and the size of the antechamber area of ​​each component).

[0301] Step S102 includes:

[0302] The lower left corner coordinates and orientation of the common elevator, the lower left corner coordinates and orientation of the fire elevator, and the lower left corner coordinates and orientation of the strong and weak current shaft are encoded as operators to simulate the relative position and orientation of each component in the core tube. The present invention designs a front room generation rule, which can generate the front room of each component and the shared front room according to the encoding, as one of the evaluation criteria for the subsequent optimization step.

[0303] Step S103 includes:

[0304] The NSGA-Ⅲ method is used to optimize the multi-objective task with constraints, including selection, crossover, mutation and other steps. Among them, the crossover adopts the SBX method, the crossover probability is set to 0.9, and the mutation adopts the PM method, and the mutation probability is set to 0.8. The selection strategy is the reference point-based method defined by the NSGA-Ⅲ algorithm. The constraint processing method adopts the CDP method that separates the objective and the constraint.

[0305] Step S104 includes:

[0306] Each uncontrolled solution finally generated by S103 is added with air shafts, doors and windows, etc., and after checking the constraints, the core tube plan layout is drawn.

[0307] Embodiment 23:

[0308] See also Figures 1 to 5 , a method for automatically generating the core tube plane layout of a high-rise building based on NSGA-Ⅲ, comprising the following steps:

[0309] S101. According to the core tube design experience and diversity requirements, two objective functions f1 and f2 are set, and according to the fire protection and ordinary residential design requirements, two constraints g1 and g2 are set. The multi-objective optimization task with constraints is defined as:

[0310] minf1=S traffic_core +S min_rec

[0311] maxf2=N shared_room

[0312] stg1=S overlap1 +S overlap2 +S overlap3 =0

[0313] g2=S fire_constrains =0

[0314] Among them, S traffic_core represents the actual area of ​​the core tube, S min_rec N represents the smallest rectangular area that can contain all core tube components. shared_room Indicates the number of shared front rooms, S overlap_items represents the overlapping area between components, S overlap_fr represents the overlapping area between each antechamber and the components, S overlap3 represents the overlapping area of ​​the corridor and each component, S fire_constrains It contains multiple judgment conditions, including whether the area size of the antechamber and shared antechamber of each component meets the fire protection regulations. The calculation method of g1 and g2 is: if the regulations are met, 0 is returned; otherwise, the difference between the specification value and the actual value is returned.

[0315] S102, such as Figure 2 As shown in the figure, the position coordinates and orientation of the stairs are set as fixed values, and the lower left corner coordinates of the ordinary elevator (x1, y1), the lower left corner coordinates of the fire elevator (x2, y2) and the lower left corner coordinates of the pipe shaft (x3, y3) and their respective orientations dir1, dir2, dir3 are encoded to simulate the relative position and orientation of each component in the core tube, specifically:

[0316] {x1,y1,dir1,x2,y2,dir2,x3,y3,dir3}

[0317] The upper limit of the variable range is set to [0,0,1,0,0,1,0,0,1], and the lower limit is set to [15,15,4,15,15,4,15,15,4]. Figure 3 As shown, after the operator is generated, the antechambers of each component are generated according to the current position and orientation. The rule is that if the antechambers overlap or are adjacent, the smallest rectangle that can contain these antechambers is taken as the shared antechamber.

[0318] S103, use NSGA-Ⅲ algorithm to optimize multi-objective tasks, such as Figure 4 As shown, first initialize the population, then iterate, each iteration process mainly includes the following steps:

[0319] (1) Perform crossover and mutation on the parent population. The crossover adopts the SBX method, which is an operator that simulates single-point binary crossover. Suppose two parent individuals X(x1, x2, …, x n ) and Y(y1,y2,…,y n ), then the two offspring individuals produced by the SBX method and It can be calculated by the following formula:

[0320]

[0321]

[0322] Among them, β is dynamically and randomly determined by the distribution factor η in the following way, and the random number rand∈[0,1):

[0323]

[0324] In the present invention, the crossover probability is set to 0.9, and the crossover distribution factor η is set to 20.

[0325] The mutation adopts the polynomial mutation (PM) method, from the parent individuals Produce offspring individuals The formula is: Let the random number randi ∈[0,1), β i The value of is determined by the random number rand i And the distribution factor η is determined by:

[0326]

[0327] In the present invention, the mutation probability is set to 0.8, and the distribution factor η of the mutation is set to 15.

[0328] (2) Sort the parent population and the subpopulation generated in (1) according to the dominance relationship. The sorting process is to first find the solution set that cannot be dominated by other solutions and divide it into the first layer, then find the solution set that cannot be dominated by other solutions from the remaining solution set as the second layer, and so on to divide it into different levels. The dominance relationship is determined according to the CDP principle: for each generated individual, calculate the two objective function values ​​f1, f2 and constraint violation values ​​g1, g2, compare the individuals pairwise, and for two individuals X and Y, X dominates Y if any of the following conditions are met:

[0329] ⅰ. If X, Y are both feasible solutions, f1(X)>f1(Y) and f2(X)>f2(Y).

[0330] ⅱ.X is a feasible solution, Y is an infeasible solution.

[0331] ⅲ. If X and Y are both infeasible solutions, CV(X)>CV(Y), where CV(X)=g1(X)+g2(X).

[0332] Among them, a feasible solution refers to a solution that satisfies the constraints, and an infeasible solution refers to a solution that does not satisfy the constraints.

[0333] (3) Select a new subpopulation from the solution set generated in (2). First, select from level 1 (the highest dominance level) according to the levels divided in (2). If the number of individuals k in the last Pareto level is exactly equal to the number of individuals K to be selected, then all individuals in this level are selected into the next generation population. If k exceeds K, individuals are selected according to the reference point selection strategy defined by the NSGA-III algorithm. The specific steps are as follows:

[0334] a) First, for the selection operation of the multi-objective function, standardization is required to enable individuals of different dimensions to be compared. The standardization process is divided into the following three steps:

[0335] ⅰ. Find the ideal point of the current population x. The ideal point is the minimum value of all objective functions of this generation of population, expressed as in is the minimum value of the i-th target, i∈[1,m]. Then perform a translation operation on the current population so that the ideal point becomes the origin:

[0336] ⅱ. Calculate the extreme point, which is a point that is very large in one target value and very small in other target values. Let w be the unit vector of each target function (i.e., the coordinate axis), then the extreme point is defined as:

[0337]

[0338] ⅲ. Construct a hyperplane and connect the extreme points. These lines form a hyperplane. After obtaining the hyperplane, set the reference point Z evenly on the hyperplane. r . Calculate the intercept a of the hyperplane with each coordinate axis i After that, the objective function can be normalized as:

[0339]

[0340] b) After standardization, the individuals in the population can be associated with the corresponding reference points. The specific steps are as follows:

[0341] ⅰ. Connect each reference point to the origin to obtain the reference line w.

[0342] ⅱ. Traverse each individual s in the population, find the reference point π(s) corresponding to the reference line with the closest vertical distance to the individual, and obtain the distance d between the individual and the reference point ⊥ (s,π(s)), the distance calculation formula is as follows:

[0343] d ⊥ (s,w)=||(sw T sw / ||w|| 2 )||

[0344] c) After constructing the mapping relationship between the population individuals and the reference point, the selection operation can be performed. Let the last level F l The number of K that should be selected, when the final level F l When the actual number of individuals k≤K, perform the following operations:

[0345] ⅰ. First select the reference point with the least number of associated individuals. If there are multiple reference points, randomly select one reference point and set the selected reference point as j. Set I j are all individuals corresponding to the reference point j. If I j If it is empty, reselect the reference point; if it is not empty, proceed to the next step.

[0346] ii. Let ρ j F l The number of individuals associated with reference point j in all non-dominated levels before the layer, if ρ j =0, then from I jThe individual with the smallest distance from the reference point is selected to enter the next generation, otherwise, the individual from I j An individual is randomly selected to proceed to the next generation.

[0347] ⅲ. Repeat the above steps until the number of selected individuals reaches K.

[0348] S104. For each uncontrolled solution finally generated in S103, determine whether the boundary satisfies the conditions for opening doors and windows. If so, add doors and windows. After generating the air shaft at the edge of the front room, check the constraints and output the final result.

[0349] The above experimental results prove that the NSGA-Ⅲ algorithm can quickly find a solution set that meets the constraints and has high availability. At the same time, because NSGA-Ⅲ adopts a reference point-based selection strategy, it ensures the diversity of generated results and can adapt to design requirements in different scenarios.

[0350] First, 1000 parent populations are initialized and 2000 iterations are performed. In each iteration:

[0351] Perform crossover and mutation on the previous generation population to produce 1,000 offspring individuals.

[0352] A total of 2,000 individuals from the parent and offspring generations are combined and sorted into different levels, with level 1 being the highest.

[0353] Now, select 1000 individuals from these 2000 individuals as the new offspring population. This new offspring population enters the next crossover mutation process. The selection process is: first select individuals from the population with level 1 to add to the offspring population until 1000 are selected. Assuming there are 100 individuals in the first level, all of them are selected, and 900 more need to be selected. If there are 900 individuals in the second level, all of them are selected and the process ends. If there are 1000 individuals in the second level, which exceeds the required 900, the selection is based on the reference point strategy and 900 are selected from these 1000.

[0354] After selecting 1000 individuals, the next iteration begins, and these 1000 individuals serve as the parent population for the next iteration.

[0355] The whole process ends when the number of iterations reaches 2000.

Claims

1. A method for automatically generating the core tube plane layout of a high-rise building based on NSGA-III, characterized in that: The following steps are involved: 1) Construct a multi-objective model; The objective functions f1 and f2 of the multi-objective model are as follows: my f1=S traffic_core +S min_rec (1) maxf2=N shared_room (2) In the formula, S traffic_core represents the actual area of ​​the core tube, S min_rec Indicates the minimum unit area of ​​the core tube containing all components, N shared_room Indicates the number of shared front rooms; The constraints g1 and g2 of the multi-objective model are as follows: s.t.g1=S overlap_items +S overlap_fr +S overlap_cor =0 (3) g2=S fire_constrains =0 (4) In the formula, S overlap_items represents the overlapping area between components, S overlap_fr represents the overlapping area between each antechamber and the components, S overlaD_cor represents the overlapping area of ​​the corridor and each component, S fire_constrains Represents a judgment function, which contains multiple judgment conditions; The judgment conditions include whether the area of ​​the antechamber of each component and the shared antechamber meets the fire protection regulations; 2) Encoding the position coordinates and orientation of each component in the core tube, and generating an antechamber corresponding to each component and / or an antechamber shared by multiple components according to the encoding, the steps comprising: 2.1) Construct a rectangular coordinate system and set the position coordinates and orientation of the stairs to fixed values; In the rectangular coordinate system, the coordinates of the lower left corner of the ith component are defined as (x i ,y i ), direction is dir i , i is a positive integer; 2.2) Simulate the relative position and orientation of each component in the core tube. The simulation variables are as follows: {x1,y1,dir1,x2,y2,dir2,...,x n ,y n ,dir n } (5) Where n is the total number of components; 2.3) Set the upper and lower limits of the simulation variables and generate the antechamber of each component according to the current position and orientation; If multiple antechambers overlap or are adjacent to each other, the smallest unit containing these antechambers shall be taken as the shared antechamber; 3) Using the NSGA-III algorithm to solve the position coordinates and orientation of each component to obtain an uncontrolled solution set, the steps include: 3.1) Initialize the population and the number of iterations q, and set 2p parent individuals X (x1, x2, ..., x n1 ), n1 is the number of parameters contained in the variable, p and q are both positive integers; the initial value of the number of iterations q is 1; 3.2) Let p parent individuals cross over with the remaining p parent individuals in pairs to produce 2p offspring individuals; the crossover method includes the SBX method; Two parent individuals cross to produce two corresponding offspring individuals and As shown below: In the formula, i1 is the parameter number in the individual; n1 is the number of parameters contained in the individual; Offspring and Parameters in; x i1 ,y i1 are the parent individuals X(x1, x2, ..., x n1 )、Y(y1,y2,...,y n1 ) in the parameters; Among them, the intermediate parameter β is as follows: Where η is the distribution factor; rand is a random number, rand∈[0,1); 3.3) The polynomial mutation method is used to mutate the offspring individuals to obtain the offspring individuals, as shown below: In the formula, is the i1th parameter of the mutated offspring individual; is the i1th parameter of the offspring individual before mutation; Among them, the intermediate parameter β i1 As shown below: In the formula, rand i1 is a random number, rand i1 ∈[0,1); 3.4) After merging the parent individuals and the offspring individuals, they are sorted according to the dominance relationship and divided into solution sets of different levels; 3.5) Let the number of iterations q = q + 1, select 2p individuals from the solution sets of different levels as the new parent population, and return to step 3.2) until the number of iterations q reaches the preset threshold Q, and the undominated solution set is obtained, where Q is a positive integer; The step of selecting 2p individuals from the solution sets of different levels as the new parent population includes: 3.5.1) Define the number of new parent populations to be 0, and the solution set of the current level to be the solution set that dominates the highest level; 3.5.2) Compare whether the number of individuals k in the new parent population is greater than the preset number of individuals 2p. If so, proceed to step 3.5.3); if not, proceed to step 3.5.4); 3.5.3) Selecting individuals from the solution set of the current level according to the reference point-based selection method defined by the NSGA-III algorithm, and selecting the individuals selected according to the reference point-based selection method defined by the NSGA-III algorithm into the new parent population, the steps comprising: 3.5.3.1) Standardizing the solution set of the current level, including the following steps: 3.5.3.1.1) Find the ideal point of the solution set of the current level, which is expressed as in, is the minimum value of the jth objective function of all individuals in the solution set of the current level, j∈[1,m], m is the number of objective functions; 3.5.3.1.2) Translate the objective function of each individual in the solution set of the current level so that the ideal point becomes the origin and construct a coordinate system; The j-th objective function of each individual in the solution set of the current level is as follows: Where f′ j (x) is the objective function after translation; f j (x) is the objective function before translation; 3.5.3.1.3) Calculate the extreme points of each objective function in the solution set of the current level as follows: In the formula, is the extreme point of the jth objective function in the solution set of the current level; w is the unit vector of each objective function; w j is the unit vector of the jth objective function in the solution set of the current level; 3.5.3.1.4) Construct a hyperplane based on the extreme points and set reference points evenly on the hyperplane Calculate the intercept a of the hyperplane with each coordinate axis j ; 3.5.3.1.5) Normalize the j-th objective function in the solution set of the current level as follows: In the formula, is the objective function after normalization; 3.5.3.2) Associating the standardized individuals to the corresponding reference points, including the following steps: 3.5.3.2.1) Each reference point and the corresponding origin Connect the lines to get the reference line W j ; 3.5.3.2.2) Traverse each standardized individual s, find the reference line W that is closest to individual s, and find the reference point π(s) that is closest to individual s on the reference line W, and calculate the distance d between individual s and the reference point π(s) ⊥ (s,π(s)), as follows: d ⊥ (s,π(s))=||(sW T sW / ||W|| 2 )|| (14) 3.5.3.3) Select the reference point π(s) with the least number of associated individuals. If there are multiple reference points π(s) with the least number of associated individuals, randomly select a reference point S from them and set the set of all individuals associated with the reference point S as I S ; 3.5.3.4) Judgment set I S Is it empty? If so, return to step 3.5.3.3); if not, go to step 3.5.3.5); 3.5.3.5) Set I S The number of individuals of all levels before the solution set of the current level is p S , determine the number of individuals p S = 0 is true, if so, then from the set I S The individual with the smallest distance from the reference point S is selected to enter the new parent population. If not, the individual is selected from the set I S An individual is randomly selected from the population to enter the new parent population; 3.5.3.6) Determine whether the number of individuals in the new parent population is equal to the preset number of individuals 2p. If so, construct a new parent population; if not, return to step 3.5.3.3); 3.5.4) Select all individuals in the solution set of the current level into the new parent population, and make the solution set of the next level as the solution set of the current level, and return to step 3.5.2); 4) According to the uncontrolled solution set obtained in step 3) and the antechamber obtained in step 2), a core tube floor plan is automatically generated based on a multi-objective model.

2. According to the method of automatic generation of high-rise building core tube plane layout based on NSGA-III in claim 1, it is characterized in that: The steps of merging the parent individuals and the offspring individuals and sorting them according to the dominance relationship include: 3.4.1) Calculate the objective function value f1, f2 and constraint violation value g1, g2 for each individual; 3.4.2) Compare individuals pairwise and sort them according to dominance relationship.

3. The method for automatically generating the core tube plane layout of a high-rise building based on NSGA-III according to claim 2 is characterized in that: The conditions under which individual X dominates individual Y are as follows: i. If X and Y are both feasible solutions, the objective function value f1(X)>f1(Y) and the objective function value f2(X)>f2(Y); ii.X is a feasible solution, Y is an infeasible solution; iii. If both X and Y are infeasible solutions, CV(X)>CV(Y), where CV(X)=g1(X)+g2(X), CV(Y)=g1(Y)+g2(Y); Among them, a feasible solution refers to a solution that satisfies the constraints, and an infeasible solution refers to a solution that does not satisfy the constraints.

4. The method for automatically generating the core tube plane layout of a high-rise building based on NSGA-III according to claim 1 is characterized in that: The step of automatically generating the core tube plan layout diagram based on the multi-objective model according to the uncontrolled solution set obtained in step 3) and the front room obtained in step 2) comprises: 4.1) Determine whether the boundary of the front room meets the conditions for opening doors and windows. If so, add doors and windows at the corresponding boundary of the front room, and generate air shafts at the boundary of the corridor or inside the front room; 4.2) Generate the core tube floor plan based on the constraints of the multi-objective model.

Images