Indoor decoration construction optimization method and system based on AI

Abstract

The invention relates to the technical field of indoor construction optimization, and discloses an AI-based indoor decoration construction optimization method and system, and the method comprises the following steps: S101, obtaining link feature data of M construction links and personnel feature data of N construction personnel, and respectively representing the link feature data and the personnel feature data through a first feature sequence and a second feature sequence; s102, preprocessing the first feature sequence and the second feature sequence to obtain third feature data and a fourth feature sequence; step S103, constructing a link construction period prediction model according to the third feature data and the fourth feature sequence, and outputting predicted construction periods of the N constructors for the M construction links; and S104, generating an optimal construction link process according to an improved parrot optimization algorithm in combination with the link construction period prediction model, wherein the construction link process represents a construction sequence of the M construction links. According to the method, through multi-dimensional data modeling, intelligent construction period prediction and an improved parrot optimization algorithm, high efficiency, accuracy and adaptability of construction process optimization are realized.

Description

Technical Field

[0001] The present invention belongs to the technical field of indoor construction optimization, and specifically relates to an indoor decoration construction optimization method and system based on AI. Background Art

[0002] With the rapid development of the construction industry, the complexity of interior decoration construction has gradually increased, and the construction process involves the coordination of a large number of links and the allocation of personnel. However, the current optimization of indoor construction processes still mainly relies on the experience of construction managers for planning and adjustment. This manual experience-driven approach has many disadvantages: on the one hand, the subjectivity and uncertainty of experience can easily lead to irrationality in construction plans, increase construction time and costs; on the other hand, manual scheduling is difficult to accurately handle the dependencies between construction links and the complexity of multi-personnel collaboration, and often leads to problems of uneven resource allocation or link conflicts. Summary of the invention

[0003] The present invention provides an AI-based interior decoration construction optimization method and system to solve the technical problems of strong subjectivity, low coordination efficiency and uneven resource allocation in related technologies.

[0004] The present invention provides an AI-based interior decoration construction optimization method, comprising the following steps:

[0005] Step S101, obtaining link characteristic data of M construction links and personnel characteristic data of N construction personnel, wherein the M link characteristic data and the N personnel characteristic data are respectively represented by a first characteristic sequence and a second characteristic sequence, wherein M and N are custom parameters set according to project requirements;

[0006] The mth sequence unit of the first feature sequence represents the link feature data of the mth construction link, where 1≤m≤M;

[0007] The nth sequence unit of the second feature sequence represents the personnel feature data of the nth construction worker, where 1≤n≤N;

[0008] The link characteristic data include: link type, construction material type, construction area and construction process complexity;

[0009] Personnel characteristic data include: health status, experience level, construction skills and concentration;

[0010] Step S102, preprocessing the first feature sequence and the second feature sequence to obtain third feature data and a fourth feature sequence;

[0011] Step S103, constructing a link construction period prediction model according to the third characteristic data and the fourth characteristic sequence, the link construction period prediction model inputs the third characteristic data and the fourth characteristic sequence, and outputs the predicted construction periods of N construction workers for M construction links respectively;

[0012] Step S104, generating the optimal construction link process according to the improved Parrot optimization algorithm combined with the link construction period prediction model, and the construction link process represents the construction sequence of M construction links.

[0013] Further, the first feature sequence and the second feature sequence are preprocessed to obtain the third feature data and the fourth feature sequence. The specific steps include:

[0014] Step S201, marking the feature values ​​in the first feature sequence and the second feature sequence that are beyond the range of three times the standard deviation of each other as outliers, and replacing them with the average value of the remaining normal values ​​in the corresponding sequences;

[0015] Step S202: extracting the area complex skewness feature according to the first feature sequence and concatenating it after the first feature sequence. The calculation formula of the area complex skewness feature is: ,in, represents the complex skewness characteristics of the area, i represents the construction link index, and Represent the area weight coefficient and complexity weight coefficient respectively, and They represent the construction area and construction process complexity of the i-th construction link respectively, and denote the nonlinear weight coefficients of area and complexity respectively, represents the construction condition correction factor, represents the exponential function, represents the corrected attenuation coefficient, represents the construction condition coefficient of the i-th construction link;

[0016] Step S203, extracting fatigue cumulative features according to the second feature sequence and splicing them after the second feature sequence. The calculation formula of fatigue cumulative features is: ,in, represents the fatigue accumulation characteristics, Indicates the current cumulative working hours of the construction personnel. represents the health status adjustment coefficient, W and H represent the work intensity coefficient and health status respectively, and the work intensity coefficient ranges from 0 to 10. The nonlinear weight coefficient representing the health status, represents the environmental correction factor, It represents the harsh environment amplification factor, Indicates the working environment score, which is evaluated by construction personnel and ranges from 0 to 1. Indicates the number of rest hours for construction workers;

[0017] Step S204, using a z-score normalization method to normalize the processed first feature sequence and the second feature sequence to obtain a third feature sequence and a fourth feature sequence.

[0018] Furthermore, the link duration prediction model includes: a first unit, a second unit, a splicing unit and a first classifier;

[0019] The first unit inputs the first feature sequence and outputs the first feature matrix;

[0020] The second unit inputs a second feature sequence and outputs a second feature vector;

[0021] The concatenation unit is used to expand the first feature matrix into a vector representation, and concatenate it with the second feature vector to obtain a concatenated vector;

[0022] The first classifier inputs the concatenated vector and outputs the predicted duration of N construction workers for M construction links.

[0023] Furthermore, the calculation formula of the first unit is: ;

[0024] in, represents the first characteristic matrix of the first unit output, represents the gating mechanism weight coefficient, represents the gating mechanism bias parameter, , and Respectively represent query weight parameters, key weight parameters and value weight parameters, represents the first characteristic sequence, T represents the transposition operation, represents element-wise multiplication, represents the dimension of the first feature sequence, represents a multi-layer perceptron, Represents the position bias parameter, sigmoid represents the sigmoid activation function, Softmax represents the Softmax activation function, and Tanh represents the Tanh activation function.

[0025] Further, the second unit includes N hidden units, the nth hidden unit inputs the nth sequence unit of the second feature sequence, outputs the nth hidden feature, the hidden feature output by the Nth hidden unit is represented as a second feature vector, and the calculation formula of the nth hidden unit is: ; ; ; ; ; ;

[0026] in, represents the nth sequence unit, and Represent the hidden features output by the nth and n-1th hidden units respectively, , , C, and They represent the update gate, reset gate, interaction matrix, candidate hidden state and dynamic update gate respectively. , and They represent the first weight parameter, the second weight parameter and the bias parameter corresponding to the update gate respectively, , and They represent the first weight parameter, the second weight parameter and the bias parameter corresponding to the reset gate respectively, T represents the transposition operation, represents the first weight parameter of the interaction matrix, , and They represent the first weight parameter, the second weight parameter and the bias parameter corresponding to the candidate hidden state respectively, , and They represent the first weight parameter, the second weight parameter and the bias parameter corresponding to the dynamic update gate respectively, sigmoid represents the sigmoid activation function, represents point-wise multiplication, Represents the tanh activation function.

[0027] Furthermore, the construction process is generated based on the improved Parrot optimization algorithm combined with the link duration prediction model. The specific steps include:

[0028] Step S301, initializing a population with a number of individuals P that meets the constraint conditions based on the reverse simulation method, where P is a custom parameter;

[0029] Step S302, designing and defining an objective function according to a link construction period prediction model;

[0030] Step S303, based on the objective function, calculating the fitness value of each individual in the population;

[0031] Step S304, simulating the parrot behavior of the individual and updating the characteristic vector of the individual in the population;

[0032] Step S305, add 1 to the current number of iterations to determine whether the current number of iterations reaches the maximum number of iterations. If so, output the individual code with the largest fitness value in the current population as the optimal construction process. Otherwise, return to step S303, where the maximum number of iterations is a custom parameter.

[0033] Furthermore, the specific steps of step S301 include:

[0034] Step S401, calculating the average values ​​of the predicted durations of N construction links by all construction personnel according to the link duration prediction model, and arranging the N average values ​​in descending order as the priority allocation order of the construction links;

[0035] Step S402, under the premise of satisfying the link dependency constraint condition, the construction links are sequentially allocated to the t-th time step according to the priority allocation order of the construction links, where t is the index of N;

[0036] Step S403: if the number of parallel links in the tth time step reaches the upper limit of the number of construction personnel, a new time step t-1 is started;

[0037] Step S404, repeating steps S402 and S403 until all construction links are allocated.

[0038] Furthermore, the calculation formula of the objective function is: ;

[0039] Among them, F represents the fitness value of the individual, max represents the maximum value operation, and j represents the construction link index. represents the set of links to which the first construction worker is assigned in the process matrix. It represents the predicted duration of the jth construction link by the first construction worker according to the link duration prediction model. represents the set of links to which the second construction worker is assigned in the process matrix, It represents the predicted duration of the second construction worker for the jth construction link obtained according to the link duration prediction model. represents the set of links to which the Nth construction worker is assigned in the process matrix, It represents the predicted duration of the jth construction link by the Nth construction worker according to the link duration prediction model.

[0040] Furthermore, the specific steps of step S304 include:

[0041] Step S501, simulating the process of a parrot searching for food by random flight, iteratively updating the individual process matrix, and the calculation formula for iterative update is: ,in, and They represent the process matrix of the kth individual at the bth iteration and the b+1th iteration respectively. and To independently draw two random variables from a standard normal distribution, represents the scaling factor, preferably, Set to 1.5, represents the scaling factor that controls the step size, represents the step size of the Levy flight distribution generation;

[0042] Step S502, simulating the behavior of a parrot staying in a specific area to observe the surrounding environment, and optimizing and adjusting the individual process matrix. The calculation formula for the optimization and adjustment is: ,in, Indicates the optimal solution influence coefficient, ranging from 0 to 1. represents the optimal solution of the population at the bth iteration, Represents a random number between 0 and 1. It represents the population mean influence coefficient, ranging from 0 to 1;

[0043] Step S503, simulating the process of the parrot adapting by observing unfamiliar behaviors, and performing random disturbance adjustment on the individual process matrix. The calculation formula of the random disturbance adjustment is: ,in, represents the disturbance coefficient, represents a standard Cauchy distribution.

[0044] The present invention provides an AI-based interior decoration construction optimization system, comprising:

[0045] A data acquisition module is used to acquire link characteristic data of M construction links and personnel characteristic data of N construction personnel, wherein the M link characteristic data and the N personnel characteristic data are respectively represented by a first characteristic sequence and a second characteristic sequence, wherein M and N are custom parameters set according to project requirements;

[0046] The mth sequence unit of the first feature sequence represents the link feature data of the mth construction link;

[0047] The nth sequence unit of the second feature sequence represents the personnel feature data of the nth construction worker;

[0048] The link characteristic data include: link type, construction material type, construction area and construction process complexity;

[0049] Personnel characteristic data include: health status, experience level, construction skills and concentration;

[0050] A data preprocessing module, used for preprocessing the first feature sequence and the second feature sequence to obtain third feature data and a fourth feature sequence;

[0051] A link construction period prediction module is used to construct a link construction period prediction model based on the third characteristic data and the fourth characteristic sequence. The link construction period prediction model takes the third characteristic data and the fourth characteristic sequence as input and outputs the predicted construction periods of N construction workers for M construction links respectively.

[0052] The process optimization module is used to generate the optimal construction process according to the improved Parrot optimization algorithm combined with the link construction period prediction model. The construction process represents the construction sequence of M construction links.

[0053] The beneficial effects of the present invention are as follows: the construction period prediction model constructed based on the link feature data and the personnel feature data of the present invention, combined with the dynamic weight adjustment mechanism and the activation function optimization, can capture the complex relationship between the features and significantly improve the accuracy of the construction period prediction and the robustness of the model;

[0054] The present invention adopts an improved parrot optimization algorithm, which achieves a dynamic balance between global search and local optimization through three major behavior simulations: random flight, environmental observation, and unfamiliar behavior learning, and improves the algorithm's global exploration ability and local convergence performance in the optimization of complex construction processes.

[0055] The present invention initializes the population based on the reverse simulation method to ensure that individuals meet actual constraints such as construction process dependencies, personnel parallel restrictions and unique allocation, effectively avoid the problem of infeasible initial solutions and improve optimization efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 It is a flow chart of an interior decoration construction optimization method based on AI of the present invention;

[0057] Figure 2 It is a flow chart of preprocessing the first feature sequence and the second feature sequence of the present invention. DETAILED DESCRIPTION

[0058] The subject matter described herein will now be discussed with reference to example embodiments. It should be understood that the discussion of these embodiments is only to enable those skilled in the art to better understand and implement the subject matter described herein, and the functions and arrangements of the elements discussed may be changed without departing from the scope of protection of the contents of this specification. Each example may omit, replace or add various processes or components as needed. In addition, the features described relative to some examples may also be combined in other examples.

[0059] like Figure 1-Figure 2 As shown, an AI-based interior decoration construction optimization method includes the following steps:

[0060] Step S101, obtaining link characteristic data of M construction links and personnel characteristic data of N construction personnel, wherein the M link characteristic data and the N personnel characteristic data are respectively represented by a first characteristic sequence and a second characteristic sequence, wherein M and N are custom parameters set according to project requirements;

[0061] The mth sequence unit of the first feature sequence represents the link feature data of the mth construction link, where 1≤m≤M;

[0062] The nth sequence unit of the second feature sequence represents the personnel feature data of the nth construction worker, where 1≤n≤N;

[0063] The link characteristic data include: link type, construction material type, construction area and construction process complexity;

[0064] Personnel characteristic data include: health status, experience level, construction skills and concentration;

[0065] Step S102, preprocessing the first feature sequence and the second feature sequence to obtain third feature data and a fourth feature sequence;

[0066] Step S103, constructing a link construction period prediction model according to the third characteristic data and the fourth characteristic sequence, the link construction period prediction model inputs the third characteristic data and the fourth characteristic sequence, and outputs the predicted construction periods of N construction workers for M construction links respectively;

[0067] Step S104, generating the optimal construction link process according to the improved Parrot optimization algorithm combined with the link construction period prediction model, and the construction link process represents the construction sequence of M construction links.

[0068] In one embodiment of the present invention, the link type is extracted through the construction plan, the type of required construction materials is obtained from the bill of materials, the construction area is determined through manual measurement, the complexity of the construction process is evaluated based on expert scores, the health status of the construction personnel is obtained through physical examination records, the personnel experience level is determined through the total project resume, the construction skills are evaluated through skill certificates or tests, and the concentration is obtained by taking the average of the concentration scores of other construction personnel on the construction personnel.

[0069] In one embodiment of the present invention, the link type is determined according to the actual construction situation and is represented by a real number code. The type of construction material is represented by a real number code. The complexity of the construction process includes: simple, medium and complex, and the complexity of the construction process is represented by a real number code. The health status includes: healthy, good and general, and the health status is represented by a real number code. The experience level is represented by the years of work. The construction skills are determined according to the skills required for the construction, which are represented by a unique hot code. The concentration is obtained by taking the average of the concentration scores of other construction workers for the construction worker, and the concentration score ranges from 0 to 100.

[0070] In one embodiment of the present invention, the first feature sequence and the second feature sequence are preprocessed to obtain the third feature data and the fourth feature sequence. The specific steps include:

[0071] Step S201, marking the feature values ​​in the first feature sequence and the second feature sequence that are beyond the range of three times the standard deviation of each other as outliers, and replacing them with the average value of the remaining normal values ​​in the corresponding sequences;

[0072] Step S202: extracting the area complex skewness feature according to the first feature sequence and concatenating it after the first feature sequence. The calculation formula of the area complex skewness feature is: ,in, represents the complex skewness characteristics of the area, i represents the construction link index, and Represent the area weight coefficient and complexity weight coefficient respectively, and They represent the construction area and construction process complexity of the i-th construction link respectively, and denote the nonlinear weight coefficients of area and complexity respectively, represents the construction condition correction factor, represents the exponential function, represents the corrected attenuation coefficient, represents the construction condition coefficient of the i-th construction link;

[0073] Step S203, extracting fatigue cumulative features according to the second feature sequence and splicing them after the second feature sequence. The calculation formula of fatigue cumulative features is: ,in, represents the fatigue accumulation characteristics, Indicates the current cumulative working hours of the construction personnel. represents the health status adjustment coefficient, W and H represent the work intensity coefficient and health status respectively, and the work intensity coefficient ranges from 0 to 10. The nonlinear weight coefficient representing the health status, represents the environmental correction factor, It represents the harsh environment amplification factor, Indicates the working environment score, which is evaluated by construction personnel and ranges from 0 to 1. Indicates the number of rest hours for construction workers;

[0074] Step S204, using a z-score normalization method to normalize the processed first feature sequence and the second feature sequence to obtain a third feature sequence and a fourth feature sequence.

[0075] In one embodiment of the present invention, the link duration prediction model includes: a first unit, a second unit, a splicing unit and a first classifier;

[0076] The first unit inputs the first feature sequence and outputs the first feature matrix;

[0077] The second unit inputs a second feature sequence and outputs a second feature vector;

[0078] The concatenation unit is used to expand the first feature matrix into a vector representation, and concatenate it with the second feature vector to obtain a concatenated vector;

[0079] The first classifier inputs the concatenated vector and outputs the predicted duration of N construction workers for M construction links.

[0080] In one embodiment of the present invention, the calculation formula of the first unit is: ;

[0081] in, represents the first characteristic matrix of the first unit output, represents the gating mechanism weight coefficient, represents the gating mechanism bias parameter, , and Respectively represent query weight parameters, key weight parameters and value weight parameters, represents the first characteristic sequence, T represents the transposition operation, represents element-wise multiplication, represents the dimension of the first feature sequence, represents a multi-layer perceptron, Represents the position bias parameter, sigmoid represents the sigmoid activation function, Softmax represents the Softmax activation function, and Tanh represents the Tanh activation function.

[0082] In one embodiment of the present invention, a multi-layer perceptron is used to learn the dynamic relationship between links based on the input features of the links. For example, links with larger areas may have a greater impact on the overall construction plan, and links with higher complexity may be more dependent on other links. By adding the Tanh activation function, the link duration prediction model's ability to express the complex relationship between features is enhanced, the attention weights of some irrelevant links are suppressed, and the robustness of the model is improved.

[0083] In one embodiment of the present invention, the second unit includes N hidden units, the nth hidden unit inputs the nth sequence unit of the second feature sequence, outputs the nth hidden feature, the hidden feature output by the Nth hidden unit is represented as a second feature vector, and the calculation formula of the nth hidden unit is: ; ; ; ; ; ;

[0084] in, represents the nth sequence unit, and Represent the hidden features output by the nth and n-1th hidden units respectively, , , C, and They represent the update gate, reset gate, interaction matrix, candidate hidden state and dynamic update gate respectively. , and They represent the first weight parameter, the second weight parameter and the bias parameter corresponding to the update gate respectively, , and They represent the first weight parameter, the second weight parameter and the bias parameter corresponding to the reset gate respectively, T represents the transposition operation, represents the first weight parameter of the interaction matrix, , and They represent the first weight parameter, the second weight parameter and the bias parameter corresponding to the candidate hidden state respectively, , and They represent the first weight parameter, the second weight parameter and the bias parameter corresponding to the dynamic update gate respectively, sigmoid represents the sigmoid activation function, represents point-wise multiplication, Represents the tanh activation function.

[0085] In one embodiment of the present invention, the interaction matrix is ​​used to capture the relationship between different construction worker features within the sequence unit, and the dynamic update gate is used to control the fusion ratio of the update gate and the candidate hidden state.

[0086] In one embodiment of the present invention, the construction process is generated according to the improved Parrot optimization algorithm combined with the link construction period prediction model, and the specific steps include:

[0087] Step S301, initializing a population with a number of individuals P that meets the constraint conditions based on the reverse simulation method, where P is a custom parameter;

[0088] Among them, each individual represents a construction process, and the individual is represented by the process matrix, which is expressed as: ,in They respectively represent the allocation of the first construction worker to the first construction link to the Mth construction link, They respectively represent the allocation of the first construction worker to the Nth construction worker for the first construction link, It indicates the allocation of the Nth construction worker to the Mth construction link. The element value of the matrix code is represented by 0 or the corresponding predicted construction period. 0 means unallocated, for example, The element value of is 4, which means that the fifth construction worker is assigned to the fourth construction link, and the predicted construction period is 4 hours;

[0089] The constraints include:

[0090] Link dependency constraints: There is a dependency relationship between construction links. The first link can only be started after the second link is completed.

[0091] Personnel parallel restriction constraint: At any time step, the number of all parallel execution links shall not exceed the total number of construction personnel N;

[0092] Link uniqueness constraint: Each link can only be assigned to one construction worker;

[0093] Step S302, designing and defining an objective function according to a link construction period prediction model;

[0094] Step S303, based on the objective function, calculating the fitness value of each individual in the population;

[0095] Step S304, simulating the parrot behavior of the individual and updating the characteristic vector of the individual in the population;

[0096] Step S305, add 1 to the current number of iterations to determine whether the current number of iterations reaches the maximum number of iterations. If so, output the individual code with the largest fitness value in the current population as the optimal construction process. Otherwise, return to step S303, where the maximum number of iterations is a custom parameter.

[0097] In one embodiment of the present invention, the improved parrot optimization algorithm achieves a balance between global search and local optimization by simulating parrot behavior, and combines the reverse simulation method to generate an initial population that satisfies the constraints, thereby improving the adaptability and feasibility of the algorithm and effectively avoiding the problem of infeasible solutions in population initialization; in each iteration, by introducing the behavior update guided by the objective function and the optimization adjustment of the process matrix, the algorithm can gradually converge to the individual with the largest fitness value, and combined with the dynamic jump mechanism, it improves the ability to jump out of the local optimal solution and enhances the global optimization performance of the algorithm.

[0098] In one embodiment of the present invention, the specific steps of step S301 include:

[0099] Step S401, calculating the average values ​​of the predicted durations of N construction links by all construction personnel according to the link duration prediction model, and arranging the N average values ​​in descending order as the priority allocation order of the construction links;

[0100] Step S402, under the premise of satisfying the link dependency constraint condition, the construction links are sequentially allocated to the t-th time step according to the priority allocation order of the construction links, where t is the index of N;

[0101] Step S403: if the number of parallel links in the tth time step reaches the upper limit of the number of construction personnel, a new time step t-1 is started;

[0102] Step S404, repeating steps S402 and S403 until all construction links are allocated.

[0103] In one embodiment of the present invention, the calculation formula of the objective function is: ;

[0104] Among them, F represents the fitness value of the individual, max represents the maximum value operation, and j represents the construction link index. represents the set of links to which the first construction worker is assigned in the process matrix. It represents the predicted duration of the jth construction link by the first construction worker according to the link duration prediction model. represents the set of links to which the second construction worker is assigned in the process matrix, It represents the predicted duration of the second construction worker for the jth construction link obtained according to the link duration prediction model. represents the set of links to which the Nth construction worker is assigned in the process matrix, It represents the predicted duration of the jth construction link by the Nth construction worker according to the link duration prediction model.

[0105] In one embodiment of the present invention, the specific steps of step S304 include:

[0106] Step S501, simulating the process of a parrot searching for food by random flight, iteratively updating the individual process matrix, and the calculation formula for iterative update is: ,in, and They represent the process matrix of the kth individual at the bth iteration and the b+1th iteration respectively. and To independently draw two random variables from a standard normal distribution, represents the scaling factor, preferably, Set to 1.5, represents the scaling factor that controls the step size, represents the step size of the Levy flight distribution generation;

[0107] Step S502, simulating the behavior of a parrot staying in a specific area to observe the surrounding environment, and optimizing and adjusting the individual process matrix. The calculation formula for the optimization and adjustment is: ,in, Indicates the optimal solution influence coefficient, ranging from 0 to 1. represents the optimal solution of the population at the bth iteration, Represents a random number between 0 and 1. It represents the population mean influence coefficient, ranging from 0 to 1;

[0108] Step S503, simulating the process of the parrot adapting by observing unfamiliar behaviors, and performing random disturbance adjustment on the individual process matrix. The calculation formula of the random disturbance adjustment is: ,in, represents the disturbance coefficient, represents a standard Cauchy distribution.

[0109] The present invention provides an AI-based interior decoration construction optimization system, comprising:

[0110] A data acquisition module is used to acquire link characteristic data of M construction links and personnel characteristic data of N construction personnel, wherein the M link characteristic data and the N personnel characteristic data are respectively represented by a first characteristic sequence and a second characteristic sequence, wherein M and N are custom parameters set according to project requirements;

[0111] The mth sequence unit of the first feature sequence represents the link feature data of the mth construction link;

[0112] The nth sequence unit of the second feature sequence represents the personnel feature data of the nth construction worker;

[0113] The link characteristic data include: link type, construction material type, construction area and construction process complexity;

[0114] Personnel characteristic data include: health status, experience level, construction skills and concentration;

[0115] A data preprocessing module, used for preprocessing the first feature sequence and the second feature sequence to obtain third feature data and a fourth feature sequence;

[0116] A link construction period prediction module is used to construct a link construction period prediction model based on the third characteristic data and the fourth characteristic sequence. The link construction period prediction model takes the third characteristic data and the fourth characteristic sequence as input and outputs the predicted construction periods of N construction workers for M construction links respectively.

[0117] The process optimization module is used to generate the optimal construction process according to the improved Parrot optimization algorithm combined with the link construction period prediction model. The construction process represents the construction sequence of M construction links.

[0118] The above describes an embodiment of the present embodiment, but the present embodiment is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the guidance of the present embodiment, ordinary technicians in this field can also make many forms, all of which are within the protection of the present embodiment.

Claims

1. An AI-based interior decoration construction optimization method, characterized in that: The following steps are involved: Step S101, obtaining link characteristic data of M construction links and personnel characteristic data of N construction personnel, wherein the M link characteristic data and the N personnel characteristic data are respectively represented by a first characteristic sequence and a second characteristic sequence, wherein M and N are custom parameters set according to project requirements; The mth sequence unit of the first feature sequence represents the link feature data of the mth construction link, where 1≤m≤M; The nth sequence unit of the second feature sequence represents the personnel feature data of the nth construction worker, where 1≤n≤N; The link characteristic data include: link type, construction material type, construction area and construction process complexity; Personnel characteristic data include: health status, experience level, construction skills and concentration; Step S102, preprocessing the first feature sequence and the second feature sequence to obtain third feature data and a fourth feature sequence; Step S103, constructing a link construction period prediction model according to the third characteristic data and the fourth characteristic sequence, the link construction period prediction model inputs the third characteristic data and the fourth characteristic sequence, and outputs the predicted construction periods of N construction workers for M construction links respectively; Step S104, generating the optimal construction link process according to the improved Parrot optimization algorithm combined with the link construction period prediction model, and the construction link process represents the construction sequence of M construction links.

2. The AI-based interior decoration construction optimization method according to claim 1 is characterized in that: Preprocessing the first feature sequence and the second feature sequence to obtain third feature data and a fourth feature sequence, specifically includes: Step S201, marking the feature values ​​in the first feature sequence and the second feature sequence that are beyond the range of three times the standard deviation of each other as outliers, and replacing them with the average value of the remaining normal values ​​in the corresponding sequences; Step S202: extracting the area complex skewness feature according to the first feature sequence and concatenating it after the first feature sequence. The calculation formula of the area complex skewness feature is: ,in, represents the complex skewness characteristics of the area, i represents the construction link index, and Represent the area weight coefficient and complexity weight coefficient respectively, and They represent the construction area and construction process complexity of the i-th construction link respectively, and denote the nonlinear weight coefficients of area and complexity respectively, represents the construction condition correction factor, represents the exponential function, represents the corrected attenuation coefficient, represents the construction condition coefficient of the i-th construction link; Step S203, extracting fatigue cumulative features according to the second feature sequence and splicing them after the second feature sequence. The calculation formula of fatigue cumulative features is: ,in, represents the fatigue accumulation characteristics, Indicates the current cumulative working hours of the construction personnel. represents the health status adjustment coefficient, W and H represent the work intensity coefficient and health status respectively, and the work intensity coefficient ranges from 0 to 10. The nonlinear weight coefficient representing the health status, represents the environmental correction factor, It represents the harsh environment amplification factor, Indicates the working environment score, which is evaluated by construction personnel and ranges from 0 to 1. Indicates the number of rest hours for construction workers; Step S204, using a z-score normalization method to normalize the processed first feature sequence and the second feature sequence to obtain a third feature sequence and a fourth feature sequence.

3. The AI-based interior decoration construction optimization method according to claim 1 is characterized in that: The link duration prediction model includes: a first unit, a second unit, a splicing unit and a first classifier; The first unit inputs the first feature sequence and outputs the first feature matrix; The second unit inputs a second feature sequence and outputs a second feature vector; The concatenation unit is used to expand the first feature matrix into a vector representation, and concatenate it with the second feature vector to obtain a concatenated vector; The first classifier inputs the concatenated vector and outputs the predicted duration of N construction workers for M construction links.

4. The AI-based interior decoration construction optimization method according to claim 3 is characterized in that: The calculation formula for the first unit is: ； in, represents the first characteristic matrix of the first unit output, represents the gating mechanism weight coefficient, represents the gating mechanism bias parameter, , and Respectively represent query weight parameters, key weight parameters and value weight parameters, represents the first characteristic sequence, T represents the transposition operation, represents element-wise multiplication, represents the dimension of the first feature sequence, represents a multi-layer perceptron, Represents the position bias parameter, sigmoid represents the sigmoid activation function, Softmax represents the Softmax activation function, and Tanh represents the Tanh activation function.

5. The AI-based interior decoration construction optimization method according to claim 3 is characterized in that: The second unit includes N hidden units. The nth hidden unit inputs the nth sequence unit of the second feature sequence and outputs the nth hidden feature. The hidden feature output by the Nth hidden unit is represented as the second feature vector. The calculation formula of the nth hidden unit is: ； ； ； ； ； ； in, represents the nth sequence unit, and Represent the hidden features output by the nth and n-1th hidden units respectively, , , C, and They represent the update gate, reset gate, interaction matrix, candidate hidden state and dynamic update gate respectively. , and They represent the first weight parameter, the second weight parameter and the bias parameter corresponding to the update gate respectively, , and They represent the first weight parameter, the second weight parameter and the bias parameter corresponding to the reset gate respectively, T represents the transposition operation, represents the first weight parameter of the interaction matrix, , and They represent the first weight parameter, the second weight parameter and the bias parameter corresponding to the candidate hidden state respectively, , and They represent the first weight parameter, the second weight parameter and the bias parameter corresponding to the dynamic update gate respectively, sigmoid represents the sigmoid activation function, represents point-wise multiplication, Represents the tanh activation function.

6. The AI-based interior decoration construction optimization method according to claim 1 is characterized in that: The construction process is generated based on the improved Parrot optimization algorithm combined with the link construction period prediction model. The specific steps include: Step S301, initializing a population with a number of individuals P that meets the constraint conditions based on the reverse simulation method, where P is a custom parameter; Step S302, designing and defining an objective function according to a link construction period prediction model; Step S303, based on the objective function, calculating the fitness value of each individual in the population; Step S304, simulating the parrot behavior of the individual and updating the characteristic vector of the individual in the population; Step S305, add 1 to the current number of iterations to determine whether the current number of iterations reaches the maximum number of iterations. If so, output the individual code with the largest fitness value in the current population as the optimal construction process. Otherwise, return to step S303, where the maximum number of iterations is a custom parameter.

7. The AI-based interior decoration construction optimization method according to claim 6 is characterized in that: The specific steps of step S301 include: Step S401, calculating the average values ​​of the predicted durations of N construction links by all construction personnel according to the link duration prediction model, and arranging the N average values ​​in descending order as the priority allocation order of the construction links; Step S402, under the premise of satisfying the link dependency constraint condition, the construction links are sequentially allocated to the t-th time step according to the priority allocation order of the construction links, where t is the index of N; Step S403: if the number of parallel links in the tth time step reaches the upper limit of the number of construction personnel, a new time step t-1 is started; Step S404, repeating steps S402 and S403 until all construction links are allocated.

8. The AI-based interior decoration construction optimization method according to claim 6 is characterized in that: The calculation formula of the objective function is: ； Among them, F represents the fitness value of the individual, max represents the maximum value operation, and j represents the construction link index. represents the set of links to which the first construction worker is assigned in the process matrix. It represents the predicted duration of the jth construction link by the first construction worker according to the link duration prediction model. represents the set of links to which the second construction worker is assigned in the process matrix, It represents the predicted duration of the second construction worker for the jth construction link obtained according to the link duration prediction model. represents the set of links to which the Nth construction worker is assigned in the process matrix, It represents the predicted duration of the jth construction link by the Nth construction worker according to the link duration prediction model.

9. The AI-based interior decoration construction optimization method according to claim 6 is characterized in that: The specific steps of step S304 include: Step S501, simulating the process of a parrot searching for food by random flight, iteratively updating the individual process matrix, and the calculation formula for iterative update is: ,in, and They represent the process matrix of the kth individual at the bth iteration and the b+1th iteration respectively. and To independently draw two random variables from a standard normal distribution, represents the scaling factor, preferably, Set to 1.5, represents the scaling factor that controls the step size, represents the step size of the Levy flight distribution generation; Step S502, simulating the behavior of a parrot staying in a specific area to observe the surrounding environment, and optimizing and adjusting the individual process matrix. The calculation formula for the optimization and adjustment is: ,in, Indicates the optimal solution influence coefficient, ranging from 0 to 1. represents the optimal solution of the population at the bth iteration, Represents a random number between 0 and 1. It represents the population mean influence coefficient, ranging from 0 to 1; Step S503, simulating the process of the parrot adapting by observing unfamiliar behaviors, and performing random disturbance adjustment on the individual process matrix. The calculation formula of the random disturbance adjustment is: ,in, represents the disturbance coefficient, represents a standard Cauchy distribution.

10. An AI-based interior decoration construction optimization system, characterized in that: An AI-based interior decoration construction optimization method as described in any one of claims 1 to 9 is adopted, comprising: A data acquisition module is used to acquire link characteristic data of M construction links and personnel characteristic data of N construction personnel, wherein the M link characteristic data and the N personnel characteristic data are respectively represented by a first characteristic sequence and a second characteristic sequence, wherein M and N are custom parameters set according to project requirements; The mth sequence unit of the first feature sequence represents the link feature data of the mth construction link; The nth sequence unit of the second feature sequence represents the personnel feature data of the nth construction worker; The link characteristic data include: link type, construction material type, construction area and construction process complexity; Personnel characteristic data include: health status, experience level, construction skills and concentration; A data preprocessing module, used for preprocessing the first feature sequence and the second feature sequence to obtain third feature data and a fourth feature sequence; A link construction period prediction module is used to construct a link construction period prediction model based on the third characteristic data and the fourth characteristic sequence. The link construction period prediction model takes the third characteristic data and the fourth characteristic sequence as input and outputs the predicted construction periods of N construction workers for M construction links respectively. The process optimization module is used to generate the optimal construction process according to the improved Parrot optimization algorithm combined with the link construction period prediction model. The construction process represents the construction sequence of M construction links.

Images