Efficient construction multi-element real-time dynamic regulation and control method based on intelligent algorithm

Abstract

The invention relates to the technical field of building construction regulation and control, and discloses an efficient construction multi-element real-time dynamic regulation and control method based on an intelligent algorithm, and the method comprises the steps: building a personnel database, and recording the competent work types, skill levels and working experience of personnel; according to a construction project, screening personnel based on an intelligent algorithm, and performing personnel allocation; monitoring the material inventory of the construction project in real time; acquiring data of personnel position, material supply, equipment operation state, construction progress and construction quality of the construction site; based on an intelligent algorithm, adjusting task distribution of the personnel according to the acquired personnel position state data; according to the efficient construction multi-element real-time dynamic regulation and control method based on the intelligent algorithm, human resources can be reasonably utilized, supply of construction materials is guaranteed, the construction efficiency can be effectively improved, the construction quality is guaranteed, and the construction cost is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of building construction control, and in particular to a high-efficiency construction multi-factor real-time dynamic control method based on an intelligent algorithm. Background Art

[0002] During the construction process, it is necessary to manage factors related to the construction, such as personnel, construction materials, construction equipment, construction progress and construction quality, in order to ensure the normal construction and management of the construction project.

[0003] There are some shortcomings in the traditional construction management method. On the one hand, the management of multiple factors such as personnel, materials, equipment, progress and quality in the construction process often relies on manual experience, lacks scientific and effective data support and precise control methods, and the following situations may occur: 1. It is easy to lead to unreasonable personnel allocation, with some positions being overstaffed and others being understaffed; 2. It is difficult to accurately control the supply of materials, and there may be a shortage of materials affecting the construction progress or a backlog of materials wasting funds; 3. Equipment management is not timely, and maintenance is only carried out after the failure occurs, causing construction to stagnate; 4. Progress control lacks flexibility and is difficult to respond to various emergencies; 5. Quality control mainly relies on post-event testing, and it is impossible to ensure quality in real time during the construction process.

[0004] On the other hand, as the scale of construction projects continues to expand and the complexity increases, the mutual influence between multiple factors such as personnel, materials, equipment, progress and quality is becoming increasingly significant. Personnel changes may affect material demand and equipment efficiency; material supply problems may cause progress delays and quality decline; equipment failures will have a chain reaction on personnel arrangements and construction progress. However, existing management methods are difficult to achieve coordinated management and real-time dynamic regulation of multiple factors, and cannot meet the requirements of modern construction efficiency, high quality and low cost. Summary of the invention

[0005] In order to solve the problem that the management of multiple factors in the above traditional construction often relies on manual experience, lacks scientific and effective data support and precise control means, the existing management methods are difficult to achieve collaborative management and real-time dynamic control of multiple factors, and cannot meet the requirements of high efficiency, high quality and low cost of modern construction, the present invention is implemented through the following technical solutions: A real-time dynamic control method for multiple factors of efficient construction based on intelligent algorithm, comprising: Establish a personnel database to record personnel's job qualifications, skill levels and work experience; According to the construction project, personnel are screened and assigned based on intelligent algorithms; Real-time monitoring of material inventory for construction projects; Collect data on personnel positions, material supply, equipment operation status, construction progress and construction quality at the construction site; Adjust personnel task allocation based on intelligent algorithms and collected personnel position data; Based on intelligent algorithms and collected construction progress and material supply data, the demand for construction project materials is predicted to adjust the material supply speed and the purchase volume of construction project materials; Based on intelligent algorithms and collected equipment operation status data, determine whether the equipment is likely to fail, so as to timely maintain and repair and dispatch the equipment; Based on intelligent algorithms and collected construction progress data, determine whether the actual construction progress matches the planned construction progress, so as to deploy personnel and / or increase the speed of material supply and / or dispatch equipment and / or adjust the construction sequence; Based on intelligent algorithms and according to the collected construction quality data, it is judged whether there are quality problems with the construction project materials or construction quality, so as to timely allocate construction project materials and / or adjust construction process parameters.

[0006] Furthermore, the particle swarm optimization algorithm is used to realize personnel allocation.

[0007] Furthermore, real-time monitoring of the material inventory of the construction project is carried out, specifically: Data collection cameras and material weight sensors are deployed in the material warehouse. The data collection cameras are used to collect the quantity data of the material inventory, and the material weight sensors are used to collect the quality data of the material inventory. At the same time, the material in and out data of the material warehouse are crawled in real time. Combined with the collected material inventory data information, the material inventory information of the construction project is monitored in real time, and the construction material supply data is obtained.

[0008] Furthermore, based on the particle swarm optimization algorithm and according to the collected personnel position data, the task allocation of personnel is adjusted, wherein the personnel position includes the location and construction status of the personnel.

[0009] Furthermore, based on an algorithm combining time series analysis with a neural network, and based on the collected construction progress and material supply data, the demand for construction project materials is predicted to adjust the material supply speed and the purchase volume of construction project materials.

[0010] Furthermore, we optimize the management of material inventory, rationally plan the inventory space according to the characteristics and usage frequency of different materials, and avoid material backlog and waste. Specifically: Genetic algorithms are used to optimize the management of material libraries. Through genetic algorithms, the best solution is found among many possible material inventory strategies.

[0011] Furthermore, in terms of equipment management, an algorithm based on the combination of particle swarm optimization algorithm and support vector machine is used to determine whether the equipment is likely to fail based on the collected equipment operation status data, so as to timely maintain, repair and dispatch the equipment.

[0012] Furthermore, in terms of construction progress control, an algorithm based on the combination of association rule mining algorithm and simulated annealing algorithm is used, and based on the collected construction progress data, it is determined whether the actual construction progress matches the planned construction progress, so as to deploy personnel and / or increase the material supply speed and / or dispatch equipment and / or adjust the construction sequence.

[0013] Furthermore, in terms of construction quality control, an algorithm based on the combination of principal component analysis and artificial neural network and based on the collected construction quality data is used to determine whether there are quality problems with the construction project materials or construction quality, so as to timely allocate construction project materials and / or adjust construction process parameters.

[0014] An efficient construction multi-factor real-time dynamic control system based on intelligent algorithms, including: Data collection module, used to collect data on personnel positions, material supply, equipment operation status, construction progress and construction quality at the construction site; The intelligent algorithm processing module is used to receive the data collected by the data acquisition module, identify the current construction status based on the received data, predict the problems that may occur during construction, and make optimal selections among multiple construction plans; A control execution module performs real-time dynamic control of multiple construction elements according to the construction plan provided by the intelligent algorithm processing module; The data acquisition sensor includes a positioning sensor, a material weight sensor, and an equipment monitoring sensor; the positioning sensor is placed on a person to monitor and obtain the location data of the construction personnel; the material weight sensor is used to obtain material supply data; the equipment monitoring sensor is used to obtain equipment operation status data; The data acquisition camera covers various areas of the construction site and various areas of the material warehouse, and is used to collect construction status data of construction personnel, material inventory data, construction progress data and construction quality data.

[0015] Compared with the prior art, the present invention has the following beneficial effects: 1. This efficient construction multi-factor real-time dynamic control method based on intelligent algorithms can dynamically adjust the number of personnel and skill combinations according to the different needs of construction tasks, ensuring that each construction link has the most appropriate human resource allocation, avoiding waste of manpower and shortage of key positions, and greatly improving personnel work efficiency. Through intelligent algorithms to analyze the work status and efficiency of personnel, work task allocation and time arrangements can be adjusted in a timely manner to improve overall construction efficiency.

[0016] 2. This efficient construction multi-factor real-time dynamic control method based on intelligent algorithms monitors material inventory and consumption in real time, uses intelligent algorithms to accurately predict material demand, communicates and coordinates with suppliers in advance to ensure timely and stable material supply, and avoids construction stagnation caused by material shortages. It optimizes inventory management, reduces material backlog waste, and reduces construction costs. At the same time, it dynamically adjusts the material procurement plan according to the construction progress to ensure that the material supply is perfectly matched with the construction progress.

[0017] 3. This efficient construction multi-factor real-time dynamic control method based on intelligent algorithms continuously monitors the operating status of equipment. Intelligent algorithms warn of equipment failures in advance, allowing maintenance personnel to handle them in a timely manner, reducing downtime, and timely and reasonably dispatching equipment. According to construction needs, equipment is allocated to the most needed links, giving full play to equipment performance and ensuring that construction progress is not affected by equipment problems.

[0018] 4. This efficient construction multi-factor real-time dynamic control method based on intelligent algorithms can dynamically adjust the construction sequence and resource allocation according to real-time data. When there are delays in progress, it can make decisions quickly and take effective measures such as increasing resource input and optimizing construction processes to ensure that the construction progress meets the planned requirements. It can respond to various emergencies, such as bad weather and design changes, and adjust the construction plan in time to ensure the stability and reliability of the construction progress.

[0019] 5. This efficient construction multi-factor real-time dynamic control method based on intelligent algorithms detects quality indicators in real time. The intelligent algorithm adjusts the construction process parameters in time according to the detection results to ensure that the construction quality always meets the standards and avoids the waste of time and resources caused by rework afterwards. It provides early warning of possible quality problems so that construction personnel can take timely measures to correct them and improve the stability and reliability of construction quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 It is a schematic diagram of the overall process of the real-time dynamic control system for efficient construction of multiple factors in an embodiment of the present invention; Figure 2 This is a schematic diagram of a data acquisition module in an embodiment of the present invention; Figure 3 This is a schematic diagram of the internal structure of the intelligent algorithm processing module in an embodiment of the present invention; Figure 4 A schematic diagram of personnel deployment by a control execution module in an embodiment of the present invention; Figure 5 A schematic diagram of equipment maintenance and scheduling by a control execution module in an embodiment of the present invention; Figure 6 It is a schematic diagram of optimizing the construction progress by the control execution module in an embodiment of the present invention. DETAILED DESCRIPTION

[0021] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0022] The embodiments of the efficient construction multi-factor real-time dynamic control method based on intelligent algorithm are as follows: See also Figure 1-Figure 6 , a method and system for real-time dynamic control of multiple factors in efficient construction based on intelligent algorithms, wherein the system includes: The data acquisition module uses high-precision sensors and high-resolution cameras to collect data on personnel positions, material supply, equipment operating status, personnel activities, material distribution, regional status, construction progress and construction quality at the construction site. The high-resolution camera is the data acquisition camera, and the high-precision sensors include positioning sensors, material weight sensors and equipment monitoring sensors.

[0023] The intelligent algorithm processing module is used to receive data transmitted by the data acquisition module, identify the current construction status based on the received data, predict problems that may arise during construction, and make optimal selections among a variety of construction plans to achieve the best construction effect.

[0024] The control execution module performs real-time dynamic control of multiple construction factors according to the construction plan provided by the intelligent algorithm processing module, including adjusting work task allocation, real-time ordering of material supply, and early warning of possible failures.

[0025] The specific implementation of a highly efficient construction multi-factor real-time dynamic control method based on intelligent algorithms in this system includes: Establish a personnel database to record personnel's job qualifications, skill levels and work experience to facilitate the allocation and adjustment of personnel.

[0026] According to the construction project, personnel are screened and assigned based on intelligent algorithms.

[0027] The detailed intelligent algorithm processing process is as follows: Detailed explanation of the application of genetic algorithm in real-time dynamic control of multiple construction factors (1) Determine the decision variables In the real-time dynamic control of multiple construction elements, accurately determining the construction elements that need to be controlled is the first crucial step. The construction progress is directly related to whether the project can be delivered on time. Resource allocation affects the rational use of manpower, material and financial resources during the construction process, and quality control parameters determine the quality standards of construction results. For example, the construction progress can be specifically broken down into time nodes for each construction stage; resource allocation covers aspects such as the supply quantity and time of materials, the deployment of equipment, and the arrangement of personnel; quality control parameters may include the strength standard of concrete, the flatness requirements of buildings, etc. Using these key elements as decision variables for the genetic algorithm provides a clear direction and object for subsequent optimization calculations.

[0028] (2) Coding and initial population formation 1. Encoding method selection The coding of decision variables needs to be reasonably selected according to their characteristics and algorithm requirements. Common coding methods include binary coding and real number coding. Binary coding is suitable for discrete variables, which represents the value of each variable with a binary string. For example, if you want to represent the different supply quantity levels of a certain material in construction, assuming there are 5 levels in total, you can use a 3-bit binary number to represent it, 000 represents the lowest level, and 111 represents the highest level. Real number coding is more suitable for continuous variables, such as the specific time value in the construction schedule or the amount of funds in resource allocation, which can be directly represented by real numbers.

[0029] 2. Initial population construction After determining the encoding method, construct the initial population. The initial population should have a certain diversity to cover as many potential solutions as possible. The initial individuals can be created by random generation. For example, for a problem that includes construction progress, resource allocation (materials and personnel), and quality control parameters, assuming that the construction progress has 3 possible time intervals, there are 4 different supply schemes for materials, 5 allocation modes for personnel, and 2 value ranges for quality control parameters, then an individual can be represented as a vector containing multiple elements, such as [schedule time interval encoding value, material supply scheme encoding value, personnel allocation mode encoding value, quality control parameter encoding value]. By randomly combining these encoding values, a certain number of individuals are generated to form the initial population. The size of the initial population is generally determined according to the complexity of the problem and the computing resources, usually between dozens and hundreds of individuals.

[0030] (3) Define the fitness function 1. Comprehensive consideration of goals The fitness function is a key tool for evaluating the quality of individuals in genetic algorithms. When constructing the fitness function, it is necessary to comprehensively consider multiple goals such as the lowest construction cost, the shortest construction period, and the highest quality. Construction costs include various expenses such as material procurement costs, equipment rental costs, and employee wages. The shortest construction period requires that the total construction time of the project be shortened as much as possible while ensuring the construction quality. The highest quality requires meeting relevant quality standards and specifications, such as the strength and stability of the building structure, the aesthetics and durability of the decoration project, etc.

[0031] 2. Function form setting A common way to construct a fitness function is to combine multiple objectives by weighted summation or product. For example, the fitness function can be set as a calculation method that comprehensively considers multiple objectives, in which cost, duration and quality are processed and combined accordingly. The determination of weights needs to be adjusted according to the actual needs and focus of the project. If the project has high requirements for cost control, the weight value of the corresponding cost can be relatively large; if time is tight, the weight of the duration should be increased; for projects with strict quality requirements, the weight of quality needs to be appropriately increased. In addition, the cost, duration and quality can also be normalized according to the specific situation so that they are calculated at the same order of magnitude to ensure the rationality and effectiveness of the fitness function.

[0032] (4) Genetic manipulation and evolution 1. Select an operation The purpose of the selection operation is to select individuals with higher fitness from the current population so that they have the opportunity to enter the next generation of population. Common selection methods include roulette selection and tournament selection. Roulette selection calculates the proportion of the individual on the roulette wheel according to the individual's fitness value. The higher the fitness, the greater the proportion, and the higher the probability of being selected. For example, suppose there are 5 individuals in the population, and their fitness values ​​are different. By calculating their proportions on the roulette wheel, and then randomly rotating the roulette wheel, the selected individual is determined. Tournament selection randomly selects several individuals from the population for comparison, and selects the individual with the highest fitness to enter the next generation. For example, 3 individuals are randomly selected each time, their fitness is compared, and the individual with the highest fitness is selected. Repeat this process many times until a sufficient number of individuals are selected to form the next generation of population.

[0033] 2. Crossover Operation Crossover is an important way to generate new individuals in genetic algorithms. For different coding methods, the crossover method is also different. In binary coding, single-point crossover or multi-point crossover is often used. Single-point crossover randomly selects a crossover point and exchanges the gene fragments of the two parent individuals after this point. For example, there are two parent individuals. Assuming that the crossover point is at a certain position, two new individuals are obtained after crossover. In real number coding, linear combination crossover or intermediate recombination crossover are generally used. For example, for two parent individuals, linear combination crossover randomly generates a crossover coefficient and calculates the new individual. The probability of the crossover operation is generally between 0.6 and 0.9, and needs to be adjusted according to the characteristics of the problem and experimental results to balance the exploration ability and convergence speed of the algorithm.

[0034] 3. Mutation Operation The purpose of mutation operation is to increase the diversity of the population and prevent the algorithm from falling into the local optimal solution. In binary coding, mutation usually changes the 0 on the gene bit to 1 or 1 to 0. The mutation probability is small, generally between 0.01 and 0.1. For example, for an individual, assuming that the mutation bit is a certain bit, a new individual is obtained after mutation. In real number coding, mutation can be a random perturbation of the gene value within a certain range. For example, for an individual, set the mutation range, randomly generate a mutation value, and then a new individual is obtained after mutation. The probability and range of the mutation operation need to be adjusted appropriately according to the specific problem. Too large a probability may destroy the existing excellent solution, and too small a probability may cause the algorithm to fail to fully explore the solution space.

[0035] 4. Multi-generational evolution process After selection, crossover and mutation operations, a new generation of population is obtained. Then these operations are repeated to continuously evolve the population. In each generation of evolution, the optimal individual of the population and its fitness value are recorded. As the number of generations increases, the population gradually converges to a better solution space. When the preset maximum number of generations is reached or other termination conditions are met (such as the fitness value reaches a certain threshold or the fitness value changes little for several consecutive generations), the algorithm stops running. The optimal individual obtained at this time is the optimal or near-optimal combination of construction elements that meets the construction requirements. This combination contains information such as the reasonable arrangement of the construction schedule, the optimal allocation of resources, and the optimal setting of quality control parameters, which can provide scientific guidance and decision-making basis for actual construction.

[0036] Application of neural network algorithm in dynamic control of multiple factors in construction: (1) Data collection and preparation 1. Data Type and Source A large amount of construction data is collected extensively, and these data cover many aspects of historical construction projects. Progress data includes the start time, end time, and comparison of the actual construction period with the planned construction period of each construction stage. For example, for a construction project, the start date of the foundation construction stage is [specific date 1] and the end date is [specific date 2]. The difference between the actual construction period and the planned construction period can reflect the speed of progress in that stage. Resource consumption data includes the amount of materials used, the consumption rate, the changes in material demand at different stages, the man-hours invested, the changes in the number of personnel in different types of work, the operating time of equipment, energy consumption, etc. Taking a large-scale construction project as an example, in the main construction stage, the monthly cement usage has a specific value, and the consumption rate will fluctuate with the progress of the construction. The material demand in different months will also change due to different construction sites and project quantities. At the same time, the number of personnel in different types of work in this stage will be adjusted according to the construction tasks, and the equipment operating time and energy consumption will also be recorded accordingly. Quality inspection data involves the quality inspection results of each link in the construction process, such as the strength test value of concrete, the dimensional deviation of the building, and the stability assessment of the structure. There are specific test values ​​for concrete strength, specific deviation values ​​for building size deviation, and professional evaluation methods and results for structural stability. These data mainly come from the project record files of the construction company, the records of on-site monitoring equipment, and the reports of the quality inspection department. The project record files of the construction company record the data of the entire project in detail, the on-site monitoring equipment records the equipment operation and material consumption in real time, and the quality inspection department detects the construction quality through professional means and forms a report.

[0037] 2. Data collation and preprocessing It is an essential step to organize and preprocess the collected data. First, the data should be cleaned to remove outliers and erroneous data. For example, the extremely high or extremely low outliers that suddenly appear in the material consumption data need to be verified. If it is a data entry error, it should be corrected; if the change in usage is caused by special circumstances, it needs to be further analyzed and marked. Then, the data is normalized to unify the data of different dimensions to the same scale range for the convenience of neural network training. For example, the number of personnel, material usage and other values ​​are normalized to the [0,1] interval, and a suitable normalization method can be used. In addition, the data needs to be divided into training set, validation set and test set. Generally, it is divided according to a certain ratio. The training set is used for network training to let it learn the data rules, the validation set is used to adjust the network parameters and evaluate the model performance during training, and the test set is used to finally evaluate the generalization ability of the trained network.

[0038] (2) Neural network structure design 1. Network type selection Multilayer Perceptron (MLP) is a common neural network structure, which is suitable for processing the complex relationship between multiple construction elements and construction results. It consists of an input layer, multiple hidden layers, and an output layer. The number of input layer nodes is determined according to the number of construction elements, such as the number of personnel, material supply (indicated by indicators such as available material inventory level and supply frequency), equipment operation status (such as equipment failure rate, operation efficiency, etc.), and other factors. If there are multiple construction elements, the number of input layer nodes corresponds to them. For example, considering the types of personnel, types of materials, and types of equipment as construction elements, assuming that there are multiple types of work, different types of materials, and several types of equipment, the number of input layer nodes is determined according to the sum of the number of these factors. The number of layers and nodes in the hidden layer is usually determined according to the complexity of the problem and experience, and the appropriate structure can be found through multiple trials and adjustments. For complex construction projects, more hidden layers and nodes may be required to capture the complex relationships in the data. Convolutional neural networks (CNNs) have advantages in processing data with spatial structures (such as construction site image data, etc.), and can be considered if the construction data contains such image information. CNN extracts and processes features of image data through structures such as convolutional layers, pooling layers, and fully connected layers, and then outputs relevant construction prediction or evaluation results. For example, in the convolutional layer, the convolution kernel slides on the image to extract local features, the pooling layer reduces the resolution of the feature map to reduce the amount of data, and the fully connected layer integrates the features and outputs the prediction results, such as the evaluation grade of the building's appearance quality.

[0039] 2. Determine the number of nodes The number of input layer nodes is determined by the number of construction elements. The number of hidden layer nodes can be preliminarily estimated through empirical formulas, and then adjusted through experiments to obtain better performance. Assume that the initial value of the number of hidden layer nodes is estimated based on the number of input and output layer nodes, and then try different numbers of hidden layer nodes in the experiment, observe the performance of the network on the training and validation sets, and select the node number configuration with better performance. The number of output layer nodes is determined according to the target to be predicted or evaluated. If it is a construction progress prediction, the output layer nodes can represent the percentage of construction completion or the remaining construction period, etc.; if it is a quality assessment result, the output layer nodes can represent the quality level (such as excellent, good, qualified, unqualified, etc., which can be represented by corresponding digital codes) or specific quality indicator values.

[0040] (3) Training and optimization 1. Principle of back propagation algorithm The back propagation algorithm is the core algorithm for training neural networks. During training, the training sample data is first input into the neural network, and the network output result is obtained through forward propagation calculation. Then the network output is compared with the actual value to calculate the error. The error is transmitted from the output layer to the input layer through back propagation. During the transmission process, the weights and biases of each layer of the network are adjusted according to the error. For each connection weight, it is adjusted according to its contribution to the error so that the network output gradually approaches the actual value. The adjustment range is controlled by the learning rate. If the learning rate is too large or too small, it will affect the training effect. Common error calculation methods include mean square error, cross entropy error, etc. For example, for mean square error, the accuracy of network prediction is measured by calculating the relationship between the number of samples, the actual value and the network output value.

[0041] 2. Training process and parameter adjustment At the beginning of training, the weights and biases of the network are randomly initialized. Then, through multiple iterations of training samples, the weights and biases are continuously adjusted. The current error is calculated for each iteration, and the network parameters are updated according to the back-propagation algorithm. During training, the error changes should be monitored. If the error no longer decreases significantly after a certain number of iterations, the learning rate or network structure may need to be adjusted. Adaptive learning rate adjustment methods such as Adagrad, Adadelta, and Adam can be used. They can automatically adjust the learning rate according to the error changes to improve training efficiency and convergence performance. At the same time, regularization methods (such as L1 regularization and L2 regularization) can be used to prevent overfitting, that is, adding a penalty term for the size of the weight to the error function so that the network does not rely too much on certain specific features or weights during the learning process, thereby improving the generalization ability of the network. During the training process, the regularization parameters need to be adjusted to balance the model complexity and generalization ability. The appropriate parameter value can be selected through methods such as cross-validation.

[0042] (4) Application and dynamic regulation 1. Real-time data input and prediction evaluation The trained neural network can be used for dynamic control of the actual construction process. Real-time collected construction data, such as the current number of personnel, material supply (real-time inventory of materials, arrival time of new materials, etc.), and equipment operating status (real-time equipment failure rate, operating parameters, etc.) are input into the neural network. After forward propagation calculation, the network quickly obtains the prediction and evaluation results of the construction status. For example, the real-time prediction of the current construction progress (such as the estimated time to complete the project according to the current status) and the real-time evaluation of the construction quality (such as whether the current construction quality meets the standards, whether there are potential quality problems, etc.) are obtained. Assume that in a construction project, the relevant data is collected in real time and input into the network, and the network predicts the possible delay of the project and the quality risks of a certain construction link.

[0043] 2. Dynamic control decisions based on results According to the prediction and evaluation results of the neural network, make corresponding dynamic control decisions on construction elements. If the construction progress forecast shows that the construction period may be delayed, measures such as increasing personnel input, speeding up material supply, and optimizing equipment scheduling can be taken. For example, personnel can be deployed from other processes to the delayed process, emergency material procurement can be arranged in advance, equipment operation time can be adjusted, and maintenance can be arranged. If the quality assessment results are not ideal, it may be necessary to adjust the construction process, increase the frequency of quality inspections, and conduct quality training for personnel. For example, adjust the construction method for the quality problem parts, increase the number of inspections, and organize personnel training. By continuously inputting real-time data and adjusting according to the results, the construction process can be made more efficient and stable to ensure the realization of construction goals.

[0044] Application of simulated annealing algorithm in construction element optimization: (1) Definition and initialization 1. State space determination The state space of construction elements covers the range of parameter values ​​of multiple key aspects of the construction process. In terms of construction progress, the state space can include the possible range of values ​​for the start and end times of each construction phase. For example, for the foundation construction phase of a building project, its start time may be within [specific time interval 1, such as the 5th-15th day] after the project starts, which depends on the preliminary preparations for the project and the deployment of related resources. The end time may be within [reasonable construction period interval, such as 20-30 days] after the start time, taking into account the process requirements of the foundation construction, the efficiency of the construction team, and the weather that may be encountered. In terms of resource allocation, the state space involves the allocation of resources such as manpower, materials, and equipment. In terms of manpower, the state space can be a combination range of the number of workers in different types of work, such as the number of carpenters between [minimum value 1, such as 10 people, maximum value 1, such as 30 people], which depends on factors such as the complexity of the construction task and the construction area; the number of steel workers between [minimum value 2, such as 15 people, maximum value 2, such as 40 people], etc., which should be determined according to the workload and construction difficulty of the steel engineering. In terms of materials, it includes the supply quantity and time range of various materials. For example, the monthly supply of a key material (such as cement) is between [lower limit, such as 200 tons, upper limit, such as 500 tons], which needs to take into account the construction progress requirements, material storage conditions and supplier supply capacity. In terms of equipment, it involves the value range of parameters such as equipment deployment and usage time. For example, the usage time of a specific construction equipment (such as a tower crane) is within [minimum time, such as 6 hours, maximum time, such as 12 hours] per day, which needs to be determined based on factors such as the urgency of the construction task, the equipment maintenance plan and the space limitations of the construction site.

[0045] 2. Objective function setting The objective function is used to measure the pros and cons of the construction plan, and is usually constructed based on key factors such as construction cost and construction period. The construction cost objective function can include the sum of various expenses such as material procurement cost, equipment rental and maintenance cost, and labor cost. For example, the material procurement cost is calculated based on the unit price and usage of different materials. Assuming that the unit price of cement is 500 yuan per ton and the usage is a certain number of tons, the cement procurement cost is the corresponding amount; the equipment rental cost is calculated according to the equipment rental time and rental rate. For example, if the tower crane rental rate is 1,000 yuan per day and the rental days are a certain number of days, the tower crane rental cost is the corresponding amount; the labor cost is calculated based on the wage standard and working hours of different types of work. The carpenter wage standard is 300 yuan per person per day and the working hours are a certain number of days, then the carpenter labor cost is the corresponding amount, etc., and these costs are added together to obtain the total construction cost. The construction period objective function focuses on the completion time of the project, and uses the deviation between the actual construction period and the planned construction period as a measurement indicator. The smaller the deviation, the closer the construction period target is to the ideal state. A penalty function for construction period deviation can be set. For example, when the actual construction period exceeds the planned construction period by a certain number of days, the penalty cost is a larger amount to reflect the severity of the delay. It is also possible to comprehensively consider the construction cost and construction period to construct a multi-objective objective function. For example, different weights are given to the construction cost and construction period through weighted summation. The weight value is determined according to the focus of the project to reflect the different emphasis on cost and construction period.

[0046] 3. Initial state selection The algorithm's search process starts from a random initial state. For the construction progress, the initial state can randomly generate the start and end time of each construction stage, and randomly select values ​​within the determined state space. For example, assuming that the start time state space of the foundation construction stage is the 5th to 15th day, a random number generator is used to randomly select an initial value, such as starting on the 8th day; the end time state space is the 28th to 38th day, and a random end value is selected, such as ending on the 32nd day. For resource allocation, the initial state can randomly determine the number of workers for different types of work, the initial supply plan for materials, and the initial allocation plan for equipment. For example, the number of carpenters is randomly determined to be a value between 10 and 30, assuming it is 20 people; the number of steel workers is a value between 15 and 40, assuming it is 30 people, etc. The initial supply plan for materials can randomly arrange the supply at each time point to take values ​​within the allowable range, such as the supply of cement in the first month is randomly determined to be 250 tons. The initial allocation plan for equipment can randomly determine the use time and order of equipment in different construction processes, etc. Assume that the daily use time of the tower crane in the foundation construction stage is randomly determined to be 8 hours.

[0047] (2) State transfer and evaluation 1. Neighborhood search Based on the current state, the neighborhood state is generated in a certain way. For the construction progress, the neighborhood state can be obtained by adjusting the start time or end time of the construction phase. For example, based on the schedule of the current basic construction phase, the start time is advanced or postponed [a certain number of days, such as 3-5 days]. If the current start time is the 8th day, then the 5th day or the 11th day can be tried as the new start time, and the end time is adjusted accordingly. At the same time, the rationality of the construction logic and resource supply must be ensured to avoid conflicts in construction procedures and excessive idleness or shortage of resources. For resource allocation, the neighborhood state can be generated by changing the number of personnel, the amount of material supply, or the equipment deployment plan. For example, increase or decrease the number of workers in a certain type of work [within a certain range of numbers, such as 5-10 people]. Assuming that the current number of carpenters is 20, you can try to adjust it to 15 or 25, etc.; adjust the supply time or supply quantity of a certain material [within a certain range of numbers, such as a supply fluctuation of 20%-30%]. If the current monthly supply of cement is 250 tons, it can be adjusted to 200 tons or 300 tons, etc., and the supply time can be advanced or postponed [within a certain time range, such as 5-10 days]; change the order of use or duration of equipment in different processes [within a certain time range, such as an increase or decrease in the equipment usage time of 1-3 hours], etc. For example, adjust the usage time of the tower crane in a certain construction process from 8 hours to 7 hours or 9 hours, or change its usage order between different processes, so as to explore the impact of different resource allocation plans on construction goals.

[0048] 2. Calculation and comparison of objective function values Calculate the objective function values ​​of the current state and the newly generated neighborhood state. If the objective function value of the new state is better than the current state, that is, the construction cost corresponding to the new state is lower or the construction period is shorter (or the comprehensive objective function value is smaller), then directly accept the new state as the current state for the next iteration. If the objective function value of the new state is worse, that is, the cost is higher or the construction period is longer (or the comprehensive objective function value is larger), then accept the new state with a certain probability. The calculation of this probability is usually based on the temperature parameter in the simulated annealing algorithm and the change in the objective function value. At the beginning of the algorithm, the temperature is high, and the probability of accepting a worse new state is relatively large, which helps the algorithm to explore in a larger search space. As the temperature gradually decreases, the probability of accepting a worse new state decreases, and the algorithm gradually converges to a better solution.

[0049] 3. Temperature update strategy The temperature update strategy has an important impact on the performance of the algorithm. Common temperature update methods include exponential decay, that is, the temperature is updated according to a certain decay coefficient, the decay coefficient is less than 1, and the temperature decreases after each iteration. Linear decay can also be used, that is, the temperature decreases by a fixed value each time. There are other more complex temperature update strategies, which can be selected and adjusted according to specific problems and experimental results. The frequency of temperature update can be to update the temperature once every certain number of iterations, or to dynamically adjust according to other rules to balance the exploration and convergence speed of the algorithm.

[0050] (3) Termination conditions and result output 1. Termination condition setting The termination condition is used to determine when the algorithm stops searching. Common termination conditions include reaching a certain number of iterations, such as setting the algorithm to stop after running 1,000 or 5,000 iterations. When the number of iterations reaches the preset value, it is considered that the algorithm has conducted enough searches and may have found a better solution or a solution close to the optimal solution. Another termination condition is that the objective function value no longer improves significantly. A threshold can be set. When the change in the objective function value is less than the threshold after multiple consecutive iterations, the algorithm is considered to have converged to a better solution and stops searching. Other conditions such as time limits can also be combined as termination conditions to ensure that the algorithm ends within a reasonable time.

[0051] 2. Determination and output of optimal solution When the algorithm meets the termination conditions, the final state is the optimal combination of construction elements. For the construction schedule, the output optimal solution includes the optimal start and end time arrangements for each construction stage, ensuring that the project can reduce costs or improve efficiency as much as possible while meeting the construction period requirements. For resource allocation, the output optimal solution clarifies the optimal number of personnel for different types of work, the optimal supply plan for materials, and the optimal deployment plan for equipment. These optimal combinations of construction elements can serve as an important basis for construction planning, guiding resource allocation and schedule scheduling in the actual construction process, so as to achieve goals such as reducing construction costs, shortening construction periods, and ensuring construction quality, and improve the overall benefits and management level of construction projects.

[0052] It should be further explained that the particle swarm optimization algorithm is used to realize personnel allocation. The detailed steps of personnel allocation using this intelligent algorithm are as follows: 1. Problem modeling and parameter definition 1. Particle definition: Each possible personnel allocation scheme is regarded as a particle. The position of the particle represents the allocation of personnel in different construction tasks or construction areas. For example, the position of a particle can be expressed as [personnel 1 is assigned to task A, personnel 2 is assigned to task B, personnel 3 is assigned to task C, ...], where personnel 1, personnel 2, etc. are specific construction personnel selected from the personnel database, and task A, task B, etc. represent different work tasks or construction areas in the construction project.

[0053] 2. Definition of fitness function: The fitness function is used to measure the quality of each particle (personnel allocation plan). According to the purpose of the invention, the fitness function should comprehensively consider factors such as construction efficiency, construction quality and construction cost.

[0054] (1) Construction efficiency: It is evaluated by calculating the estimated completion time of each task. The estimated completion time is estimated based on the skill level and work experience of the personnel and the complexity of the task. For example, for a task that requires a higher skill level, if a person with rich experience and high skills is assigned, the estimated completion time will be shorter. Construction efficiency can be measured by the sum of the estimated completion time of all tasks. The smaller the sum, the higher the construction efficiency.

[0055] (2) Construction quality: Evaluation indicators are set based on the impact of personnel skill levels and work experience on construction quality. For example, for some key tasks with high quality requirements, assigning personnel with relevant professional skills and rich experience can improve construction quality. A quality weight can be set for each task, and a quality score can be calculated based on the personnel allocation. The higher the quality score, the better the construction quality.

[0056] (3) Construction cost: Consider the salary cost of personnel and possible additional costs caused by unreasonable personnel allocation, such as overtime costs. The salary cost is determined based on the skill level of the personnel and the market salary level, and the additional cost is estimated based on the urgency of the task and the matching degree between personnel allocation and task requirements. The lower the construction cost, the better the solution.

[0057] Taking the above factors into consideration, the fitness function can be defined as: Fitness = α×(1 / construction efficiency)+β×construction quality+γ×(1 / construction cost), where α, β, and γ are weight coefficients determined according to the actual needs of the project, which are used to balance the importance of construction efficiency, construction quality, and construction cost in fitness evaluation.

[0058] (II) Initializing the particle swarm 1. Generate initial particles: randomly generate a group of particles, that is, the initial personnel allocation plan. Randomly select personnel from the personnel database and assign them to various construction tasks or construction areas to ensure that the allocation plan represented by each particle is feasible, that is, it meets the basic requirements of the construction task, such as the number of personnel requirements, skill requirements, etc.

[0059] 2. Set initial velocity: Assign an initial velocity to each particle. The velocity dimension is the same as the particle position dimension, indicating the adjustment direction and amplitude of the personnel allocation plan. The initial velocity can be randomly generated within a reasonable range, such as between [-v_max, v_max], where v_max is the maximum velocity value determined by experience or experiment, ensuring that the particle has a certain search range and flexibility in the initial stage.

[0060] 3. Iteration process 1. Fitness evaluation: For each particle, its fitness value is calculated according to the defined fitness function. The fitness value reflects the comprehensive performance of the personnel allocation scheme represented by the particle in terms of construction efficiency, construction quality and construction cost.

[0061] 2. Update individual optimality and global optimality: (1) Individual optimal update: Each particle will record its optimal position so far (i.e., the personnel allocation plan with the best fitness value) during the iteration process, which is called the individual historical optimal position (pBest). If the fitness value of the current particle is better than its previously recorded individual optimal fitness value, the individual optimal position is updated to the current position.

[0062] (2) Global optimal update: In the entire particle swarm, find the particle with the best fitness value, and its position is called the global optimal position (gBest). The global optimal position represents the optimal personnel allocation solution found among all particles so far.

[0063] 3. Speed ​​and position update: (1) Speed ​​update: The particle adjusts its speed based on its own historical best position (pBest) and the global best position (gBest) to search in a better direction. When updating the speed, the influence of the particle's previous speed, as well as the difference between its own best position, the global best position and the current position, should be considered. The direction and magnitude of the speed adjustment should be determined according to certain rules.

[0064] (2) Position update: adjust the position of the particle according to the updated speed, that is, the personnel allocation plan. During the update process, it is necessary to ensure the rationality and feasibility of personnel allocation. For example, if a task requires personnel with specific skills and the updated allocation plan does not meet this requirement, appropriate adjustments or corrections are required.

[0065] 4. Constraint processing: In the process of personnel allocation, there may be some constraints, such as some personnel can only engage in specific types of work or tasks, or some tasks have clear requirements on the number of personnel, etc. After each update of the particle position (personnel allocation plan), it is necessary to check whether these constraints are met. If not, appropriate measures need to be taken to correct them. For example, by reallocating personnel who do not meet the constraints, or adjusting the speed and position of the particles, it can continue to search for a better solution while meeting the constraints.

[0066] (IV) Termination Condition Judgment The iteration process is repeated until a pre-set termination condition is met. The termination condition can be one of the following: 1. Reach the maximum number of iterations: Set a maximum number of iterations N_max. When the number of iterations reaches N_max, the algorithm stops running and outputs the currently found optimal personnel allocation plan (global optimal position gBest).

[0067] 2. The fitness value meets the requirements: If the fitness value of the global optimal particle reaches or exceeds a preset threshold, it means that a better personnel allocation plan that meets the requirements has been found and the algorithm terminates.

[0068] 3. The change amplitude is less than the threshold: Calculate the change amplitude of the global optimal fitness value in several consecutive iterations. If the change amplitude is less than a given threshold, it means that the algorithm has converged to a relatively stable state, and continued iteration may not bring significant improvement. At this time, the algorithm can be stopped.

[0069] (V) Output the optimal personnel allocation plan When the algorithm meets the termination condition, the personnel allocation plan represented by the global optimal position (gBest) is output, which is the final optimal personnel allocation plan. This plan can guide the allocation of personnel in the construction project to achieve the goals of rationally utilizing human resources, improving construction efficiency, ensuring construction quality and reducing construction costs. At the same time, during the construction process, the algorithm can be re-run according to the actual situation (such as personnel changes, construction task adjustments, etc.), and the personnel allocation plan can be adjusted dynamically in real time to adapt to the changing needs of the construction project.

[0070] Real-time monitoring of the material inventory of the construction project; specifically, deploying data acquisition cameras and material weight sensors in the material warehouse, using the data acquisition cameras to collect the quantity data of the material inventory, using the material weight sensors to collect the quality data of the material inventory, and at the same time crawling the material in and out data in the material warehouse in real time, combined with the material inventory data information collected by the acquisition module, real-time monitoring of the material inventory information of the construction project, and then obtaining the construction material supply data. The combination of the material warehouse in and out data and the particle swarm optimization algorithm is as follows: 1. Data collection and preprocessing 1. Real-time data collection: Establish a connection with the material warehouse management system, and crawl the material in and out data in real time through data interfaces or sensors and other technical means. These data include detailed information such as the type, quantity, in and out time, and supplier information of the materials. For example, every time a material enters the warehouse, record the name of the incoming material (such as steel, cement, wood, etc.), specification model, in-warehouse quantity, in-warehouse time, and supplier name and contact information; every time a material leaves the warehouse, record the relevant information of the outgoing material and the construction task or location used.

[0071] 2. Data preprocessing: Clean and organize the collected raw inbound and outbound data to remove noise data and outliers. For example, if the material inflow quantity recorded at a certain time is negative or obviously inconsistent with the actual situation (such as far exceeding the normal purchase volume), the data needs to be verified and corrected. At the same time, the data is classified and formatted so that it can be effectively combined with intelligent algorithms later. For example, different types of material data are stored separately and sorted in chronological order to prepare for data analysis and application.

[0072] 2. Application in Particle Swarm Optimization Algorithm 1. Fitness function adjustment: Inclusion of the timeliness of material supply: When defining the fitness function, consider the impact of material supply on construction. Incorporate the timeliness of material supply as an important factor into the fitness evaluation. For example, if the supply of a key material is delayed during construction, it may hinder the construction progress, thereby increasing construction costs and extending the construction period. Therefore, in the fitness function, a penalty term related to the delay in material supply can be set. If the actual warehousing time of a certain material is later than the planned warehousing time, the fitness value will be deducted accordingly based on the length of the delay and the importance of the material to the construction. On the contrary, if the material can be put into storage on time or in advance, and the quantity meets the construction needs, a certain amount of bonus points can be given in the fitness value.

[0073] Consideration of material inventory cost factors: In addition to the timeliness of supply, material inventory cost is also an important consideration. Too high material inventory will occupy a lot of funds and increase costs; while too low inventory may lead to construction interruptions and affect construction efficiency. In the fitness function, the inventory level and inventory cost of materials are considered comprehensively. Relevant indicators can be constructed based on factors such as material inventory turnover, storage cost, and out-of-stock cost. For example, define an inventory cost coefficient that is related to the average inventory level and inventory cost of the material. If the inventory management of the material in the construction plan represented by the particle can make the inventory cost coefficient low, that is, the inventory level is reasonable and the cost is low while ensuring the normal progress of construction, then the fitness value will increase accordingly.

[0074] 2. Impact of particle position and velocity updates: Adjustment of personnel tasks based on material inventory: The position of the particle indicates the allocation of personnel in different construction tasks or construction areas. When the material inventory situation changes, for example, some materials are in excess inventory and some key materials are in short supply, the construction tasks and personnel allocation need to be adjusted accordingly. If there is too much inventory of a certain material, more personnel may need to be assigned to construction tasks related to the material to speed up the consumption of inventory; if the key material inventory is insufficient, personnel may need to be adjusted to prioritize tasks that do not depend on the material, or the construction schedule may need to be adjusted to wait for material replenishment. In this case, the position of the particle (personnel allocation plan) will be adjusted accordingly according to the material inventory situation. The speed update will also be affected. For example, if the personnel allocation needs to be adjusted quickly to cope with changes in material inventory, the speed adjustment may increase, allowing the particle to move to a new and better position faster.

[0075] Coordinated optimization of material supply and construction progress: In the process of speed and position update, the coordinated relationship between the material supply rhythm and the construction progress is considered. By analyzing the material in and out data and the construction plan, the use of personnel and equipment is reasonably arranged to ensure that the construction progress matches the material supply. For example, if it is predicted that the supply of a certain key material will decrease in the future, then in the particle update process, the proportion of personnel allocated to the construction tasks related to the material can be appropriately reduced, and the construction speed can be adjusted to avoid construction stagnation due to material shortages. Conversely, if the material supply is sufficient and the construction progress is lagging behind, the particle update speed can be accelerated and the personnel input of related construction tasks can be increased to speed up the construction progress.

[0076] 3. Real-time dynamic adjustment and feedback mechanism 1. Real-time monitoring and adjustment: During the construction process, the in-and-out data of the material warehouse and the actual progress of the construction are continuously monitored in real time. According to the real-time data, the particle swarm optimization algorithm is constantly adjusted and optimized. For example, if it is found that the actual material usage is not in line with expectations, or there are new changes in the supply of materials (such as delayed delivery by suppliers, increased material loss at the construction site, etc.), the fitness function is immediately recalculated, and the speed and position of the particles are updated to adjust the personnel allocation and construction plan.

[0077] 2. Establishment of feedback mechanism: Establish a feedback mechanism to compare and analyze the actual situation during the construction process with the optimization results of the intelligent algorithm. If it is found that the actual construction effect deviates from the expected goal, timely feedback will be given to the intelligent algorithm so that the parameters and strategies of the intelligent algorithm can be adjusted. For example, if the construction progress is still lagging behind or the cost is not effectively controlled after the personnel allocation plan and construction plan optimized by the intelligent algorithm are implemented, the reasons can be analyzed through the feedback mechanism. It may be that the weights of certain factors in the fitness function are set unreasonably, or the intelligent algorithm is not accurate enough when dealing with certain constraints. According to the feedback information, the intelligent algorithm is improved and optimized accordingly to continuously improve the accuracy and adaptability of the intelligent algorithm and realize the real-time dynamic and precise control of multiple construction factors.

[0078] Through the above method, the real-time crawled material library in-and-out data is closely combined with the particle swarm optimization algorithm, which can realize the comprehensive optimization and dynamic regulation of personnel allocation, material supply and other construction elements in the construction process, and improve the overall management level and efficiency of the construction project.

[0079] Collect data on personnel position, material supply, equipment operating status, construction progress and construction quality at the construction site, where personnel position includes personnel location and construction status; specifically, deploy data collection cameras at the construction site, place positioning sensors on construction workers, and deploy equipment monitoring sensors in equipment at the construction site.

[0080] The data collection cameras at the construction site are used to collect the construction status data, construction progress data, and construction quality data of the construction workers. The positioning sensors are used to monitor and obtain the location data of the construction workers in real time. The equipment monitoring sensors are used to obtain the data on the equipment operation status.

[0081] Among them, the equipment includes quality inspection equipment, which is used to detect the quality of construction materials and construction quality, so as to obtain construction quality data.

[0082] In terms of personnel management during construction, personnel task allocation is adjusted based on intelligent algorithms and collected personnel position data.

[0083] It should be further explained that according to the construction plan and task requirements, after the task is issued, the intelligent algorithm is used to screen out suitable personnel from the personnel database, and after reasonable allocation, the construction personnel enter the construction site to carry out construction work. The data collection cameras and positioning sensors at the construction site monitor the positions of the personnel. If it is found that the personnel in a certain area are too concentrated or idle, the task allocation of the personnel will be adjusted in time to improve the utilization rate of human resources.

[0084] Specifically, the particle swarm optimization algorithm is used to realize the comprehensive control of personnel allocation and multiple construction factors, including personnel adjustment. The particle swarm optimization algorithm has obvious advantages in dealing with the optimization problem of multiple construction factors, because it can quickly find the optimal personnel adjustment plan according to the real-time construction situation and constraints to ensure efficient construction. The detailed steps of personnel adjustment using this intelligent algorithm are as follows: 1. Problem modeling and parameter definition 1. Particle definition: The personnel adjustment plan is regarded as a particle. The position of the particle represents the change in the allocation status of personnel in the construction task, area or stage. For example, the position of a particle may be [personnel 1 is transferred from task A to task B, personnel 2 increases the working time in task C, personnel 3 is transferred from area D to area E, etc.], personnel are individuals in the construction team, and tasks, areas, etc. represent different units or positions of the construction.

[0085] 2. Fitness function definition: The fitness function is used to measure the quality of the personnel adjustment plan, taking into account factors such as construction efficiency, quality, cost, and material supply.

[0086] (1) Construction efficiency: Evaluate the impact of personnel adjustments on the progress of construction tasks. Calculate efficiency indicators based on the expected completion time and actual progress of tasks, and set weights for different tasks for comprehensive calculation. If the completion time of tasks is advanced after personnel adjustments, construction efficiency will be improved.

[0087] (3) Construction quality: The skills and experience of personnel affect the construction quality. When adjusting personnel, ensure that there are skilled workers in key links, evaluate quality indicators based on skill matching and quality history data, and improve quality, which will increase the quality score.

[0088] (3) Construction cost: Personnel adjustment involves costs such as wages, training and overtime. Calculate the total cost under the adjustment plan. The lower the cost, the higher the adaptability.

[0089] Matching of material supply and personnel: Consider the coordination between material inventory and supply rhythm and personnel adjustment. If the personnel adjustment can better match material use and supply to avoid construction delays caused by backlogs or shortages, the adaptability will be increased.

[0090] (II) Initializing the particle swarm 1. Generate initial particles: Randomly generate the initial personnel adjustment plan. Perform random perturbations based on the current personnel allocation, such as randomly swapping personnel tasks or areas, changing working hours, etc., to ensure the diversity of initial particles and cover a wider solution space.

[0091] 2. Set initial speed: Assign an initial speed to the particle. The speed dimension is the same as the position, representing the amplitude and direction of personnel adjustment. Set it randomly within a reasonable range, and determine the upper and lower limits of the speed according to the common amplitude and frequency of personnel adjustment, so that the particle has the ability to explore.

[0092] 3. Iteration process 1. Evaluate fitness: For each particle's personnel adjustment plan, calculate its value according to the fitness function. By obtaining data such as construction task progress, quality inspection reports, cost accounting, and material inventory in and out records, comprehensively evaluate the pros and cons of the plan.

[0093] 2. Update individual optimality and global optimality: (1) Individual optimal update: Each particle records its own optimal position (pBest) and updates if the current fitness value is better.

[0094] (2) Global optimal update: Find the particle with the best fitness value in the particle swarm, and its position is the global optimal (gBest).

[0095] 3. Speed ​​and position update: (1) Speed ​​update: Particles adjust their speed based on their own historical optimal position and the global optimal position. When the gap with the individual optimal position is large, the speed is adjusted in the direction of the individual optimal position; when a better solution is found in the global optimal position, the speed direction and magnitude are adjusted by attraction. At the same time, it is restricted by constraints to ensure that personnel adjustments are reasonable and feasible.

[0096] (2) Position update: Update the particle position according to the speed, that is, update the personnel adjustment plan. The update must meet the actual construction requirements and constraints, such as skill requirements, working time restrictions, etc. If the constraints are violated, corrections must be made, such as reallocating personnel or adjusting speed.

[0097] 4. Dynamic adjustment considering real-time factors: (1) Changes in construction progress: Monitor the progress in real time. If the task is delayed or advanced, adjust the particle speed and position according to the deviation. If a key task is delayed, increase the input of relevant task personnel or adjust the configuration. During the update process, focus on improving the personnel adjustment direction of the task construction efficiency.

[0098] (2) Changes in material supply: Based on the material in and out data, adjust the personnel plan when there is a shortage or surplus. If there is a shortage of materials, reduce the number of people who are responsible for tasks that depend on the materials or transfer them to other tasks. Consider the material supply factor when updating particles, so that personnel can be adjusted to adapt to the material situation.

[0099] (3) Emergency handling: For emergencies such as staff absence and equipment failure, timely adjust algorithm parameters and constraints. Recalculate the feasibility and adaptability of the personnel adjustment plan, adjust the speed and position, and find the optimal strategy under the existing personnel conditions.

[0100] (IV) Termination Condition Judgment Iterate until the termination condition is met. This includes reaching the maximum number of iterations, the fitness value being stable or meeting the requirements, the end of the construction project phase, or changes in conditions. The maximum number of iterations is determined by the complexity of the project and computing resources; convergence is considered when the fitness value changes slightly or reaches the threshold; reinitialization and iteration may be required when the project phase changes or the conditions change significantly.

[0101] (V) Output the optimal personnel adjustment plan When the particle swarm optimization algorithm meets the termination conditions, it outputs the personnel adjustment plan represented by the global optimal position as the basis for adjusting the construction personnel. During construction, the situation changes are continuously monitored and the algorithm optimization plan is re-run when necessary to ensure that the construction proceeds smoothly under the optimal personnel configuration.

[0102] In terms of material supply management, based on intelligent algorithms and according to the collected construction progress and material supply data, the demand for construction project materials is predicted to adjust the material supply speed and the purchase volume of construction project materials.

[0103] It should be further explained that data collection cameras and material weight sensors are used to monitor the material inventory in the material warehouse in real time to obtain construction material supply data. Intelligent algorithms predict future material demand based on the material consumption rate in the construction progress data and material supply data. When the material inventory in the material warehouse is lower than the set level (safety warehouse level), purchase orders are sent to suppliers in a timely manner to ensure timely supply of materials.

[0104] Specifically, a combination of time series analysis and neural network algorithms is used to predict future material demand. This combined algorithm can fully utilize the characteristics of time series data and the nonlinear fitting ability of neural networks to accurately capture the complex relationship between material demand and influencing factors, provide a scientific basis for the supply of construction materials, ensure smooth construction, and avoid cost increases and construction delays. The detailed steps for using this intelligent algorithm to predict future materials are as follows: 1. Data collection and organization 1. Time series data collection: Collect historical data on material usage, such as usage, purchase volume, inventory level and corresponding time, such as monthly steel usage, cement purchase time and quantity in the past year.

[0105] 2. Data collection of relevant factors: Determine the factors that affect material demand, such as construction schedule, construction process changes, seasonal factors, market price fluctuations, etc., and integrate them with material data. For example, record the time and key nodes of each stage of the construction project, and clarify the impact of seasons on material demand and material price changes.

[0106] 3. Data cleaning and preprocessing: Clean the data to remove outliers and erroneous data, and use methods such as mean or median substitution to correct them. Normalize the data to make its value range uniform, which is convenient for the intelligent algorithm to calculate and analyze.

[0107] 2. Time series analysis 1. Trend analysis: Use moving average method, exponential smoothing method, etc. to analyze the historical data trend of material usage. Moving average method highlights long-term trends by calculating the average value of data within a certain time window, such as using 3-month moving average to analyze the trend of steel usage. Exponential smoothing method predicts future trends based on different weights, and selects appropriate smoothing coefficients to process material usage to obtain trend forecast values.

[0108] 2. Seasonal analysis: Identify seasonal fluctuations in material demand and use seasonal index methods to analyze. If it is found that material demand increases during the peak construction period in summer, calculate the quantitative changes in the summer seasonal index and make seasonal adjustments to the data to predict seasonal changes.

[0109] 3. Feature Engineering and Data Preparation 1. Feature construction: Construct neural network input features based on time series and related factor data. In addition to the original data, new features are derived, such as time periodicity features (using the sine and cosine values ​​of the month to represent the seasonal cycle), construction progress stage features, price change rate, etc., to help the neural network learn the rules.

[0110] 2. Dataset division: Divide the data into training set, validation set and test set. The training set accounts for 60%-70% for training the model, the validation set accounts for 20%-30% for adjusting hyperparameters and evaluating performance, and the test set accounts for 10%-20% for evaluating the generalization ability of the model. Ensure that the data set is randomly divided and contains data from various situations.

[0111] (IV) Neural network construction and training 1. Network structure design: Build a suitable neural network structure, such as selecting RNN and its variants (LSTM, GRU, etc.) considering the characteristics of time series. For example, build an LSTM network with input layer, hidden layer and output layer. The number of input layer nodes is determined according to the number of features, the number of hidden layer layers and nodes is adjusted according to the complexity of the data and computing resources, and the number of output layer nodes is the predicted future material demand.

[0112] 2. Model training: Use the training set data to train the neural network, and adjust the network weights and biases through the back-propagation algorithm to minimize the error between the predicted value and the actual value. Select a suitable loss function (such as MSE, MAE), use stochastic gradient descent or its improved algorithm to optimize the parameters, set appropriate hyperparameters such as learning rate, batch size, and number of training rounds, and use the validation set to monitor model performance to prevent overfitting.

[0113] 5. Model evaluation and optimization 1. Selection of evaluation indicators: Use MSE, MAE, MAPE, RMSE and other indicators to comprehensively evaluate model performance. MAPE can intuitively reflect the relative error between the predicted value and the actual value.

[0114] 2. Model evaluation and analysis: Use the test set data to evaluate the model, input the feature data into the model to obtain the predicted value and compare it with the actual value to calculate the evaluation index. Analyze the evaluation results to find out the advantages and disadvantages of the model. For example, in some cases, the prediction error may be large because the data pattern has not been learned or there are factors that have not been considered.

[0115] 3. Model optimization: Optimize the model according to the evaluation results. You can adjust the network structure, increase or decrease the number of hidden layers and nodes, etc. If the amount of data is small, you can use data enhancement technology; introduce new features or factors, such as policy indicators; try to integrate learning methods to combine multiple models to improve stability and accuracy.

[0116] (VI) Future material demand forecast and application 1. Predict future material demand: Use the optimized model to predict future material demand. It can predict demand at different time granularities (day, week, month, quarter, etc.) and provide a reference for material procurement, inventory management and production planning.

[0117] 2. Application of results and decision support: Apply the forecast results to construction material management decisions, arrange procurement plans reasonably to ensure supply, optimize inventory management to reduce costs, adjust the construction schedule according to the forecast, ensure the smooth progress of the project, achieve efficient coordination and dynamic regulation of multiple construction elements, and improve project benefits and management levels. At the same time, the model needs to be continuously monitored and updated to ensure the accuracy of the forecast.

[0118] At the same time, the material inventory is optimized and managed. According to the characteristics and usage frequency of different materials, the inventory space is reasonably planned to avoid material backlog and waste. Specifically, the genetic algorithm is used to achieve optimized management of the material library. Genetic algorithm is an intelligent search algorithm based on the principles of natural selection and genetic variation. It has the advantages of strong global search ability and good adaptability. It is very suitable for solving complex combinatorial optimization problems in material library management. During the construction process, the optimized management of the material library is crucial to ensure the construction progress, reduce costs and improve resource utilization. Through genetic algorithms, the best or better solutions can be found among many possible material inventory strategies to achieve efficient management of the material library. The detailed steps for optimizing the management of the material library using this intelligent algorithm are as follows: 1. Problem Modeling and Coding 1. Determine the decision variables: Identify the decision variables that need to be optimized. In material library management, these variables may include the purchase quantity, inventory level, replenishment time point, etc. of different materials. For example, for a variety of commonly used building materials, such as steel, cement, wood, etc., determine the purchase quantity and inventory control parameters of each material in different time periods as decision variables.

[0119] 2. Encoding method selection: Use appropriate encoding methods to convert decision variables into chromosome forms that can be processed by genetic algorithms. Common encoding methods include binary encoding, real number encoding, etc. For material library management problems, real number encoding may be more appropriate because decision variables such as purchase quantity and inventory level are usually continuous values. For example, a chromosome can be expressed as [steel purchase quantity 1, steel inventory upper limit 1, steel replenishment time point 1, cement purchase quantity 2, cement inventory upper limit 2, cement replenishment time point 2, ...], where each value corresponds to a specific value of a decision variable.

[0120] 2. Initializing the population 1. Generate initial individuals: According to the coding method, a certain number of initial individuals are randomly generated, and each individual represents a possible material inventory management plan. The generation of initial individuals should be carried out within a reasonable range, taking into account constraints such as the actual demand for materials, the supply capacity of suppliers, and the storage capacity of warehouses. For example, the purchase volume of steel cannot exceed the estimated demand of the construction project in a certain period of time, and the inventory limit must be less than the actual storage capacity of the warehouse.

[0121] 2. Construct the initial population: Combine the generated initial individuals into an initial population. The size of the population is determined by the complexity of the problem and the computing resources. Generally speaking, the larger the population, the higher the possibility that the genetic algorithm will search for the global optimal solution, but the computing cost will also increase accordingly. Usually, the size of the initial population can be selected between dozens and hundreds of individuals.

[0122] (III) Fitness function design 1. Determine the objective function: define the objective function used to measure the quality of the material inventory management solution. The objective function usually includes multiple considerations, such as minimizing costs and maximizing service levels. In the optimization management of the material library, the cost aspects can include procurement costs, inventory holding costs, and out-of-stock costs. Procurement costs are related to the purchase price and purchase volume of materials; inventory holding costs are related to inventory levels and holding time; out-of-stock costs depend on the frequency of out-of-stocks and the impact of out-of-stock quantities on construction. The service level can be measured by the timeliness rate of meeting construction needs, that is, the proportion of required materials that can be provided in a timely manner.

[0123] 2. Fitness function construction: The objective function is converted into a fitness function. The larger the value of the fitness function, the better the material inventory management plan corresponding to the individual. A common method is to transform the objective function appropriately so that its value is as large as possible. For example, if the goal is to minimize the total cost, the fitness function can be defined as the inverse of the total cost. For the service level, a reward factor can be set according to the timeliness rate of meeting the demand. When the timeliness rate is higher, the fitness value increases accordingly.

[0124] 4. Genetic manipulation 1. Selection operation: According to the fitness value of the individual, a part of individuals are selected from the current population as parents to generate the next generation of individuals. Common selection methods include roulette selection and tournament selection. Roulette selection determines the probability of being selected based on the proportion of the individual's fitness value in the total fitness of the entire population. The higher the fitness value, the greater the probability of being selected. Tournament selection randomly selects a certain number of individuals from the population to compete, and the individual with the highest fitness value is selected as the parent.

[0125] 2. Crossover operation: Perform a crossover operation on the selected parent individuals to generate new offspring individuals. The crossover operation simulates the gene recombination process in biological genetics, generating offspring individuals with different characteristics by exchanging some gene fragments of the parent individuals. For the material library management problem, single-point crossover, multi-point crossover and other methods can be used. For example, in a single-point crossover, a crossover point is randomly selected, and the gene fragments of the two parent individuals after this point are exchanged. The crossover probability is usually selected between 0.6 and 0.9, which determines the frequency of the crossover operation.

[0126] 3. Mutation operation: Perform mutation operation on newly generated offspring individuals to increase the diversity of the population. Mutation operation simulates the phenomenon of gene mutation in biological inheritance, and generates new individual characteristics by randomly changing certain gene values ​​on individual chromosomes. In material library management, mutation can make small random adjustments to decision variables such as purchase volume and inventory level. The probability of mutation is generally small, usually between 0.01 and 0.1, to avoid excessive damage to excellent individuals.

[0127] 5. Iteration process 1. Population update: The offspring individuals generated by genetic manipulation replace some individuals with lower fitness in the current population to form a new population. In this way, the individuals in the population gradually evolve in a better direction.

[0128] 2. Repeat genetic operations: Repeat selection, crossover and mutation operations on the new population, and continuously iterate and optimize. In each iteration, record the best individual in the current population and its fitness value.

[0129] 3. Termination condition judgment: Set the termination condition of the iteration, and stop the algorithm when the condition is met. The termination condition can be reaching the predetermined maximum number of iterations, the fitness value reaching the preset optimal value, or the change in the fitness value in several consecutive iterations is less than a certain threshold. For example, if the maximum number of iterations is set to 100 times, when the algorithm executes to the 100th iteration, the iteration will be stopped regardless of whether the optimal solution is found. Alternatively, if the fitness value is very close to the ideal optimal value, or the change in the fitness value in 10 consecutive iterations is very small, it can also be considered that the algorithm has converged and the iteration is stopped.

[0130] 6. Output the optimal solution When the genetic algorithm meets the termination condition, the individual with the highest fitness value in the population is output as the optimal solution for optimizing the management of the material library. This solution includes the values ​​of decision variables such as the optimal purchase quantity of each material, the inventory limit, and the replenishment time point. Managing the material library according to this solution can reduce costs, improve service levels, and effectively utilize resources. In the actual construction process, as the construction progress advances and the external environment changes, the genetic algorithm can be re-run regularly to dynamically adjust and optimize the material library management solution to adapt to changing needs.

[0131] Through the above steps, the genetic algorithm can effectively realize the optimized management of the material library and improve the material management efficiency and economic benefits of the construction project.

[0132] In terms of equipment management, based on intelligent algorithms and according to the collected equipment operation status data, it is determined whether the equipment is likely to fail, so as to carry out timely maintenance, repair and dispatch of equipment.

[0133] It should be further explained that by installing equipment monitoring sensors on the equipment, continuously monitoring the operating status parameter data of the equipment, and using intelligent algorithms to analyze the operating parameters of the equipment, such as the operating temperature, pressure, vibration, etc. of the equipment, early warning of equipment failures. When abnormal equipment operation is detected and a failure may occur, maintenance personnel will promptly inspect and repair it. At the same time, according to the needs of the construction task, the equipment is reasonably dispatched. For example, during the concrete pouring stage, equipment such as concrete pump trucks are preferentially allocated to the construction area to improve the efficiency of equipment use. Establish an equipment maintenance plan, and perform regular maintenance and maintenance according to the equipment's usage time and operating status to extend the service life of the equipment.

[0134] Specifically, the particle swarm optimization algorithm (PSO) is combined with the support vector machine (SVM) to realize the allocation, fault warning and / or scheduling of equipment. PSO is good at searching for the optimal solution and is suitable for decision optimization of equipment allocation and scheduling; SVM performs well in classification prediction and can be used for equipment fault warning. The combination of the two can improve the efficiency and reliability of equipment management. The detailed steps for using this intelligent algorithm for equipment allocation, fault warning and / or scheduling are as follows: 1. Equipment Allocation - Based on Particle Swarm Optimization (PSO) 1. Problem modeling and coding Determine the decision variables for allocating equipment to construction tasks or areas, and use integer encoding to represent the particle position vector, that is, the equipment allocation plan.

[0135] 2. Initialize particle swarm Randomly generate the initial device allocation scheme as particles, considering the reasonable allocation of the number of devices and tasks. Set the initial velocity vector for the particles.

[0136] 3. Design fitness function The fitness function is constructed by comprehensively considering factors such as construction efficiency (task completion time), cost (operation, transportation, etc.) and equipment utilization.

[0137] 4. Particle update and iteration Update particle speeds, adjust equipment allocation plans based on actual factors, and make corrections if they are unreasonable. Calculate fitness, update individual optimal and global optimal positions, and stop when termination conditions (such as maximum number of iterations or fitness convergence) are reached.

[0138] 5. Output allocation plan: Get the optimal plan for equipment allocation.

[0139] 2. Equipment Failure Warning - Based on Support Vector Machine (SVM) 1. Data collection and preprocessing Collect equipment operating parameters, maintenance and fault history information. Label the data into two categories: normal and faulty, clean and pre-process, remove outliers and noise, and perform standardization or normalization.

[0140] 2. Feature selection and extraction Analyze and find out the features related to the fault, such as temperature, vibration, etc. If necessary, use methods such as principal component analysis to reduce the dimension and extract the main information.

[0141] 3. SVM model training Divide the training set and test set. Select appropriate kernel function and parameters, and train the model with features as input and fault labels as output.

[0142] 4. Model evaluation and optimization Use indicators such as accuracy and recall rate to evaluate. If the results are not ideal, adjust feature processing, parameters, or add data for optimization.

[0143] 5. Application warning: Real-time data is input into the model to predict failures and issue warnings.

[0144] 3. Equipment Scheduling - Based on the Combination of Particle Swarm Optimization Algorithm and Genetic Algorithm 1. Problem modeling and coding The decision variables for equipment scheduling, such as task arrangement and movement path, are determined using hybrid coding.

[0145] 2. Initialize the population In PSO, the randomly generated equipment scheduling scheme is particles, and the genetic algorithm randomly generates chromosomes. Set relevant parameters.

[0146] 3. Design fitness function The fitness function is constructed by considering task completion time, device idle time, moving distance and cost.

[0147] 4. Algorithm iteration and operation PSO updates speed and position, corrects unreasonable solutions, and updates the optimal position. Genetic algorithm performs selection, crossover, and mutation operations. The two algorithms are operated alternately and stop when the termination condition is reached.

[0148] 5. Output scheduling plan: Output the optimal plan for equipment scheduling management Through the application steps of the above intelligent algorithms, effective management of construction equipment can be achieved, the efficiency and quality of construction projects can be improved, and adjustments and optimizations can be made according to actual conditions.

[0149] In terms of construction progress control, based on intelligent algorithms and according to the collected construction progress data, it is determined whether the actual construction progress matches the planned construction progress, so as to deploy personnel and / or increase the material supply speed and / or dispatch equipment and / or adjust the construction sequence.

[0150] It should be further explained that the data collection cameras at the construction site monitor and record the construction progress in real time, and compare the actual progress recorded with the construction plan progress. When the construction progress is found to be lagging behind, the reasons are analyzed and corresponding measures are taken. If the construction progress is delayed due to insufficient personnel, personnel can be deployed from other projects or temporary workers can be added; if the material supply is not timely, the material supply speed can be accelerated; if the equipment failure affects the construction progress, maintenance personnel are given priority to repair it, and the construction sequence is adjusted reasonably to avoid the entire project being stagnant due to equipment failure. By constantly adjusting the allocation of construction resources and optimizing the construction process, the construction progress is ensured to meet the planned requirements.

[0151] Specifically, the association rule mining algorithm is combined with the simulated annealing algorithm to realize the analysis of the causes of construction progress delay and construction optimization. The association rule mining algorithm can find the potential correlation between factors in the construction data and determine the causes of progress delay; the simulated annealing algorithm can search for the optimal construction adjustment plan in the complex solution space to achieve progress optimization. The two work together to improve the efficiency and quality of project management. The detailed steps for using this intelligent algorithm to analyze the causes of construction progress delay and optimize construction are as follows: 1. Analysis of causes of construction progress delay - based on association rule mining algorithm 1. Data collection and organization Collect construction-related data, such as task start and end time, construction period, resource input, environmental factors, construction technology, etc. Obtain from management systems, logs, monitoring equipment, etc.

[0152] Preprocess data to remove noise, errors and duplicate data, and classify and encode them for easy analysis, such as classifying tasks, quantifying resources, and digitizing environmental factors.

[0153] 2. Build a transaction database The construction tasks or stages are considered as affairs, and the related factors are considered as items. For example, "foundation construction" is a matter, and "number of personnel", "materials", "weather", "process" and so on are considered as items.

[0154] The database is constructed by organizing data by transactions and items, and each record corresponds to a transaction and related item information.

[0155] 3. Mining association rules Set support and confidence thresholds based on experience and data characteristics.

[0156] Use algorithms such as Apriori to mine association rules that meet the threshold, such as "If there are few people and the weather is bad, the probability of extending the construction period is high."

[0157] Analyze the rule results to find the combination of factors that are highly correlated with progress delays and present them in visualization or reports.

[0158] (II) Construction optimization - based on simulated annealing algorithm 1. Problem modeling and coding Determine decision variables, including task sequence adjustment, resource reallocation, process improvement, etc. For example, use numbers to encode task sequences, and real numbers or combinations to encode resource adjustments.

[0159] Select an appropriate coding method to construct a chromosome representation scheme, such as "[task sequence coding, personnel number adjustment value, equipment usage time adjustment value, etc.]".

[0160] 2. Initialization solution The initial feasible plan is randomly generated based on the current situation and constraints to ensure compliance with task logic and resource constraints.

[0161] The target value of the initial solution is calculated based on the objective function, which covers aspects such as shortening the construction period, reducing costs, and improving quality.

[0162] 3. Set simulated annealing parameters Determine the initial temperature, which should be high enough to ensure search capability. It can be set empirically and then adjusted experimentally.

[0163] Set the cooling rate to control the convergence speed, generally between 0.9 and 0.99.

[0164] Set the end temperature, usually a small value such as 0.01 or 0.001.

[0165] The number of iterations at each temperature is determined, taking into account both efficiency and accuracy.

[0166] 4. Simulated annealing process Perturbing the current solution at the current temperature generates a new solution, such as exchanging the order of tasks and adjusting the number of resources.

[0167] Calculate the target value of the new solution and compare it with the current solution.

[0168] The probability of accepting a new solution is determined based on the Metropolis criterion. If the new solution is superior, it is accepted directly; otherwise, it is accepted based on probability.

[0169] The temperature is reduced at the cooling rate and the cycle is repeated until the termination temperature.

[0170] 5. Output optimization results: After the algorithm is completed, the optimal solution is output, that is, the construction optimization plan, including task adjustment, resource allocation, process improvement and other measures. The construction is adjusted accordingly to solve the problem of delayed progress, and the plan can be dynamically updated according to actual changes.

[0171] In terms of construction quality control, based on intelligent algorithms and according to the collected construction quality data, it is judged whether there are quality problems with the construction project materials or construction quality, so as to timely allocate construction project materials and / or adjust construction process parameters.

[0172] It should be further explained that at the construction site, quality inspection equipment is used to detect the quality of construction materials and the quality of construction after completion. For example, the strength of concrete, the welding quality of steel bars, etc. The intelligent algorithm adjusts the construction process parameters in a timely manner according to the test results. For example, when it is detected that the concrete strength does not meet the standard, the possible reasons are analyzed, such as improper mix ratio, insufficient curing time, etc., and the concrete mix ratio is adjusted or the curing time is extended. Early warning is given to links where quality problems may occur, and quality control is strengthened. For example, during the steel bar welding process, the welding quality is monitored in real time through the detection equipment. Once problems such as loose welding are found, the construction personnel are immediately notified to make rectifications to avoid the expansion of quality problems.

[0173] Specifically, principal component analysis (PCA) and artificial neural network (ANN) are combined to achieve material quality analysis and quality problem warning. PCA can reduce the dimension of high-dimensional data of material quality, extract key features, and remove redundant information; ANN can learn rules from complex data, accurately analyze quality and predict problems. The combination of the two improves the accuracy of material quality analysis and the reliability of warning, ensures the quality of construction materials, and avoids adverse consequences caused by quality problems. The detailed steps for using this intelligent algorithm to analyze material quality and warn of quality problems are as follows: 1. Material quality data collection and preprocessing 1. Data collection: Collect material quality-related data from multiple channels, such as physical properties (density, hardness, etc.), chemical composition, production process parameters, supplier information, construction use and performance, etc. For example, the strength, composition, production batch and stress conditions of steel in construction.

[0174] 2. Data cleaning: Remove outliers and erroneous data, and use statistical methods to identify and process them. For example, based on the mean and standard deviation, verify and correct deviating data, or delete those that cannot be corrected. Check the completeness, and a small number of missing values ​​can be filled with the mean, median, or based on the prediction of related variables.

[0175] 3. Data standardization: Standardize the cleaned data, such as using the Z-score standardization method to convert the data into a distribution with a mean of 0 and a standard deviation of 1, so that data of different dimensions and magnitudes are comparable, which is conducive to subsequent algorithm calculation and analysis.

[0176] 2. Feature extraction and dimensionality reduction based on principal component analysis (PCA) 1. Calculate the covariance matrix: For the standardized data matrix X (n samples, m features), calculate the covariance matrix C=(1 / n)*X*X^T to reflect the correlation between features.

[0177] 2. Determine the principal components: Perform eigenvalue decomposition on the covariance matrix to obtain eigenvalues ​​and eigenvectors. Sort by eigenvalue from large to small, calculate the cumulative contribution rate, select the first k principal components that reach a certain threshold (such as 80%-90%), project the original data onto these principal component eigenvectors, and obtain the new data matrix Y=X*V after dimensionality reduction, remove redundant information, and retain the main components.

[0178] 3. Model construction and training based on artificial neural network (ANN) 1. Network structure design: Design a multi-layer feedforward neural network. The number of input layer nodes is equal to the number of principal components after dimensionality reduction. The number of hidden layers and nodes is selected according to the complexity of the problem. The number of output layer nodes is determined according to the task (classification or prediction).

[0179] 2. Data partitioning: Divide the reduced data into training set, validation set, and test set, usually in a ratio of 70%, 15%, and 15%, to ensure randomness and representativeness.

[0180] 3. Model training: Train the neural network with the training set, adjust the weights and biases through the back-propagation algorithm, minimize the error between the prediction and the actual value, select the appropriate error function and optimization algorithm, set the hyperparameters and adjust them through the validation set, monitor the training to prevent overfitting, and use the early stopping method or regularization method.

[0181] 4. Model evaluation and optimization: Use the test set to evaluate the model, analyze the performance based on indicators such as accuracy and root mean square error, and if it is not ideal, adjust the network structure, data enhancement, feature engineering, or use ensemble learning and other methods to optimize.

[0182] 4. Application of early warning of material quality issues 1. Real-time data processing: Material quality data is collected in real time during construction, cleaned and standardized according to the preprocessing steps, and then the principal component data is obtained through PCA dimensionality reduction.

[0183] 2. Quality analysis and early warning: Input the reduced real-time data into the trained ANN model to analyze the quality and predict problems. If the output value approaches or exceeds the early warning threshold, an early warning signal is issued.

[0184] 3. Decision-making and measures: After receiving the early warning, the management personnel take measures such as suspending use, testing and analysis, tracing the source, etc. At the same time, they adjust the construction plan, such as purchasing spare materials, speeding up the progress and reducing dependence, etc., to achieve dynamic management and prevention, ensure smooth construction and achieve quality goals, and continuously update data and optimize models to adapt to changes.

[0185] While the embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that many changes, modifications, substitutions and variations can be made to the embodiments without departing from the principles and spirit of the invention.

Claims

1. A real-time dynamic control method for efficient construction of multiple factors based on intelligent algorithm, characterized in that: include: Establish a personnel database to record personnel's job qualifications, skill levels and work experience; According to the construction project, personnel are screened and assigned based on intelligent algorithms; Real-time monitoring of material inventory for construction projects; Collect data on personnel positions, material supply, equipment operation status, construction progress and construction quality at the construction site; Adjust personnel task allocation based on intelligent algorithms and collected personnel position data; Based on intelligent algorithms and collected construction progress and material supply data, the demand for construction project materials is predicted to adjust the material supply speed and the purchase volume of construction project materials; Based on intelligent algorithms and collected equipment operation status data, determine whether the equipment is likely to fail, so as to timely maintain and repair and dispatch the equipment; Based on intelligent algorithms and collected construction progress data, determine whether the actual construction progress matches the planned construction progress, so as to deploy personnel and / or increase the speed of material supply and / or dispatch equipment and / or adjust the construction sequence; Based on intelligent algorithms and according to the collected construction quality data, it is judged whether there are quality problems with the construction project materials or construction quality, so as to timely allocate construction project materials and / or adjust construction process parameters.

2. The method for real-time dynamic control of multiple factors in efficient construction based on intelligent algorithm according to claim 1 is characterized by: Particle swarm optimization algorithm is used to realize personnel allocation.

3. The method for real-time dynamic control of multiple factors in efficient construction based on intelligent algorithm according to claim 1 is characterized by: Real-time monitoring of material inventory for construction projects, specifically: Data collection cameras and material weight sensors are deployed in the material warehouse. The data collection cameras are used to collect the quantity data of the material inventory, and the material weight sensors are used to collect the quality data of the material inventory. At the same time, the material in and out data of the material warehouse are crawled in real time. Combined with the collected material inventory data information, the material inventory information of the construction project is monitored in real time, and the construction material supply data is obtained.

4. The method for real-time dynamic control of multiple factors of efficient construction based on intelligent algorithm according to claim 1 is characterized by: Based on the particle swarm optimization algorithm and according to the collected personnel position data, the task allocation of personnel is adjusted, where the personnel position includes the location and construction status of the personnel.

5. The method for real-time dynamic control of multiple factors of efficient construction based on intelligent algorithm according to claim 1 is characterized by: Based on an algorithm that combines time series analysis with neural networks, and based on the collected construction progress and material supply data, the demand for construction project materials is predicted to adjust the material supply speed and the purchase volume of construction project materials.

6. The method for real-time dynamic control of multiple factors in efficient construction based on intelligent algorithm according to claim 5 is characterized by: Optimize the management of material inventory, reasonably plan the inventory space according to the characteristics and usage frequency of different materials, and avoid material backlog and waste. Specifically: Genetic algorithms are used to optimize the management of material libraries. Through genetic algorithms, the best solution is found among many possible material inventory strategies.

7. The method for real-time dynamic control of multiple factors in efficient construction based on intelligent algorithm according to claim 1 is characterized by: In terms of equipment management, an algorithm based on the combination of particle swarm optimization algorithm and support vector machine is used to determine whether the equipment is likely to fail based on the collected equipment operation status data, so as to ensure timely maintenance and dispatch of the equipment.

8. The method for real-time dynamic control of multiple factors in efficient construction based on intelligent algorithm according to claim 1 is characterized by: In terms of construction progress control, an algorithm based on the combination of association rule mining algorithm and simulated annealing algorithm is used to determine whether the actual construction progress matches the planned construction progress based on the collected construction progress data, so as to deploy personnel and / or increase the material supply speed and / or dispatch equipment and / or adjust the construction sequence.

9. The method for real-time dynamic control of multiple factors in efficient construction based on intelligent algorithm according to claim 1 is characterized by: In terms of construction quality control, an algorithm based on the combination of principal component analysis and artificial neural network is used to determine whether there are quality problems with the construction project materials or construction quality based on the collected construction quality data, so as to timely allocate construction project materials and / or adjust construction process parameters.

10. A high-efficiency construction multi-factor real-time dynamic control system based on intelligent algorithm, using the high-efficiency construction multi-factor real-time dynamic control method based on intelligent algorithm as claimed in any one of claims 1 to 9, characterized in that: include: Data collection module, used to collect data on personnel positions, material supply, equipment operation status, construction progress and construction quality at the construction site; The intelligent algorithm processing module is used to receive the data collected by the data acquisition module, identify the current construction status based on the received data, predict the problems that may occur during construction, and make optimal selections among multiple construction plans; A control execution module performs real-time dynamic control of multiple construction elements according to the construction plan provided by the intelligent algorithm processing module; The data acquisition sensor includes a positioning sensor, a material weight sensor, and an equipment monitoring sensor; the positioning sensor is placed on a person to monitor and obtain the location data of the construction personnel; the material weight sensor is used to obtain material supply data; the equipment monitoring sensor is used to obtain equipment operation status data; The data acquisition camera covers various areas of the construction site and various areas of the material warehouse, and is used to collect construction status data of construction personnel, material inventory data, construction progress data and construction quality data.

Images