AI-based engineering quantity list automatic compiling method and system

Abstract

The invention relates to the technical field of engineering management, in particular to an AI-based engineering quantity list automatic compiling method and system, and the method comprises the following steps: constructing a model through analyzing a construction task dependency relationship, tracking the progress in real time, and predicting changes through a neural network to generate change analysis; and optimizing delay task adjustment time and resource configuration in combination with constraint conditions, evaluating a weather risk level and dynamically adjusting a construction period, and finally dynamically updating a project quantity list based on resource and risk results. According to the method, a task network structure with a quantifiable scoring mechanism is constructed by analyzing the dependency relationship between the tasks, so that the logic association between the tasks not only stays on the static hierarchy presentation, but also introduces the difference weight of time and resource dimensions to realize dynamic modeling on the dependency of key paths; and depth support is provided for subsequent prediction and adjustment. Tracking of task progress changes does not depend on a static progress comparison mode any more, but deep extraction of actual construction data is carried out.

Description

Technical Field

[0001] The present invention relates to the technical field of project management, and particularly to an AI-based automatic preparation method and system for a bill of quantities. Background Art

[0002] The technical field of project management encompasses multiple aspects such as project planning, design, construction, and management, involving the management and coordination throughout the project lifecycle. The core elements of this field include project schedule control, cost management, quality control, risk management, resource allocation, etc., aiming to ensure the efficient and smooth completion of engineering projects.

[0003] Among them, the automatic preparation method for the bill of quantities refers to an automated technical method for the preparation of the bill of quantities during the project budget preparation process. By analyzing the design drawings and related documents of the engineering project, relevant data related to the quantities of work are automatically extracted, and a bill of quantities is automatically generated according to the specific requirements of the project.

[0004] In the existing technology for handling the preparation of the bill of quantities, it mainly relies on the static information extraction method from design drawings and related documents, lacking in-depth analysis of the logical relationships between construction tasks, resulting in the inability to promptly identify the key impact chains during task changes. For example, in a project with multiple tasks running in parallel and intersecting, a delay in one task may trigger a chain reaction, and existing means are difficult to predict and respond efficiently. During the progress analysis, it is mostly based on manual comparison and empirical judgment, with a long response cycle for sudden delays, affecting the stability of the overall project schedule plan. In terms of resource allocation, the original methods usually rely on static resource tables and fixed project schedule plans, unable to effectively adapt to the actual changing resource supply and demand situations, often resulting in resource waste or scheduling conflicts. In the face of external uncontrollable factors such as weather impacts, existing solutions mainly rely on historical experience warnings, lacking a scientific quantitative model for task-level risk assessment and adjustment guidance, resulting in an unclear floating range of the project schedule. The update of the bill of quantities usually relies on manual revision, with data updates lagging behind, easily forming a disconnection problem between plan execution. Summary of the Invention

[0005] The purpose of the present invention is to address the drawbacks existing in the prior art and propose an AI-based automatic preparation method and system for a bill of quantities.

[0006] To achieve the above objective, the present invention adopts the following technical solutions: An AI-based automatic preparation method for a bill of quantities, comprising the following steps: S1: Obtain engineering construction task information, analyze the dependencies between construction tasks, and construct a task dependency model according to the dependencies; S2: Based on the task dependency model, track the task changes in the construction progress in real time, use a neural network to predict task changes, and generate a progress change analysis result; S3: Obtain the mutual influence of task delays in the progress change analysis result, analyze the optimal adjustment time of the task duration under the condition of meeting the constraint conditions by setting the constraint conditions of engineering tasks, and make corresponding adjustments to generate a resource allocation adjustment result; S4: Analyze the potential risks of weather factors on construction tasks by setting fuzzy evaluation rules for multiple risk levels, adjust the construction period and resource allocation according to the analysis result, and generate a risk impact adjustment result; S5: Based on the resource allocation adjustment result and the risk impact adjustment result, dynamically compile and update the bill of quantities to generate a dynamically adjusted bill of quantities.

[0007] As a further solution of the present invention, the task dependency model includes task nodes, dependency paths, and critical paths. The progress change analysis result includes the predicted delay duration, task delay level, and the degree of influence on related tasks. The resource allocation adjustment result includes the construction period compression duration, resource allocation plan, and task adjustment priority. The risk impact adjustment result includes the weather risk level, affected task list, and task priority ranking. The dynamically adjusted bill of quantities includes the adjusted task duration, optimized resource allocation list, and task implementation priority order.

[0008] As a further solution of the present invention, the obtaining steps of the task dependency model are specifically as follows: S111: Based on the engineering construction task information, obtain the start time, duration, and task identification attributes of each task, extract the preceding first task and the subsequent second task that depends on the preceding first task, judge the time dependency relationship between tasks, analyze the duration data of each first task and the second task, and generate a task time sequence dependency analysis result; S112: According to the task time sequence dependency analysis result, establish a task node set and an edge set of dependency paths, draw a dependency relationship diagram between tasks, judge whether the dependency relationship between tasks meets the time sequence requirements, analyze whether there are loops between task nodes, and screen the data set to generate task dependency structure map data; S113: Based on the task dependency structure map data, extract the sequence of task nodes. By calculating the time difference between the first task and the second task nodes, combined with the maximum resource consumption and minimum resource consumption of the first task and the second task, quantitatively evaluate the dependency relationship between the two task nodes, and establish a task dependency model.

[0009] As a further solution of the present invention, the obtaining steps of the progress change analysis result are specifically as follows: S211: Obtain the construction progress data of the first task node and the second task node, including the construction period and the schedule deviation. Based on the task dependency model, integrate the progress change of the first task node as the pre-input data of the deep neural network to generate the progress change input data; S212: Input the progress change input data into the deep neural network for training, predict the start time and the change of the construction period of the second task node, calculate the new start time and the construction period of the second task node based on the progress change of the first task node, and output the task progress prediction result; S213: According to the task progress prediction result, quantify the impact of the change of the construction period of the second task node on the overall project progress, including the delay time and the delay amplitude, and generate the progress change analysis result.

[0010] As a further solution of the present invention, the step of obtaining the resource allocation adjustment result is specifically as follows: S311: Based on the progress change analysis result, collect and sort out the delay data of the first task node, including the deviation between the completion time and the planned completion time, analyze the impact of the delay of the first task on the start time of the second task node, and generate the first task delay impact data; S312: Based on the first task delay impact data, use the constraint optimization algorithm to set the conditions of the project task time constraint, resource constraint and dependency constraint, conduct constraint analysis on the construction period and resource allocation of the second task node, and according to the set conditions, adopt the formula: ; Calculate the objective function of the construction period of the second task under the satisfaction of all constraints , determine the optimal adjustment time according to the objective function, and generate the optimal construction period and resource allocation list; Wherein, is the construction period after the adjustment of the second task the value after normalization within the allowable time range, is the resource requirement after the adjustment of the second task the value after normalization within the allocable resource range, is the dependency score of the first task on the second task the value after normalization, , , : are the weight parameters of the construction period, resource and dependency respectively; S313: Based on the optimal construction period and resource allocation list, readjust the construction period and resource allocation of the second task to generate the resource allocation adjustment result.

[0011] As a further solution of the present invention, the step of obtaining the risk impact adjustment result is specifically as follows: S411: Obtain the current weather data, including precipitation, wind speed, and temperature, combine it with the progress and resource requirement information of the construction task to generate a weather data set; S412: Based on the weather data set, according to the fuzzy evaluation rules of multiple risk levels set, match the weather factors of precipitation, wind speed, and temperature with the risk levels, and use the formula: ; Calculate the fuzzy risk value affected by the weather , analyze the potential construction risks of each weather condition, and evaluate the impact degree of the weather condition on the construction task to generate a weather risk assessment result; Wherein, represents precipitation, represents wind speed, represents temperature, , , respectively represent the intermediate control values of the corresponding weather factors, , , respectively represent the maximum reference values of each factor in the set risk level; S413: According to the weather risk assessment result, adjust the construction period and resource allocation, prioritize the tasks according to the degree of influence of the tasks by the weather, and generate a risk impact adjustment result.

[0012] As a further solution of the present invention, the steps for obtaining the dynamically adjusted bill of quantities are specifically as follows: S511: Based on the resource allocation adjustment result and the risk impact adjustment result, call the construction period adjustment amount, resource configuration change value, and priority order in each task node, perform unified numbering association on the corresponding task entries to generate a task adjustment parameter structure; S512: According to the task adjustment parameter structure, match the construction period value, resource data, and priority order of each task to the field positions associated with the corresponding task numbers in the bill of quantities in sequence, and mark the corresponding fields in the original bill of quantities as updatable according to the field positions to generate list field update data; S513: Call the list field update data, perform data replacement operations on the fields that need to be updated in the original bill of quantities structure, and reassign the list version structure after the replacement is completed. The numbering is represented by the current date plus the version number, covering the display fields of the original list file, and generating a dynamically adjusted bill of quantities.

[0013] An AI-based automatic compilation system for bill of quantities, where the AI-based automatic compilation system for bill of quantities is used to execute the above-mentioned AI-based automatic compilation method for bill of quantities. The system includes: The task dependency modeling module obtains the engineering construction task information, analyzes the dependency relationship between construction tasks, and constructs a task dependency model according to the dependency relationship; The progress prediction and change analysis module, based on the task dependency model, tracks the task changes in the construction progress in real time, uses a neural network to predict task changes, and generates a progress change analysis result; The task duration optimization module obtains the mutual influence of task delays in the progress change analysis result, analyzes the optimal adjustment time of the task duration under the condition of meeting the constraint conditions by setting the constraint conditions of the engineering tasks, and makes corresponding adjustments to generate a resource allocation adjustment result; The risk assessment and impact adjustment module analyzes the potential risks of weather factors on construction tasks by setting fuzzy assessment rules for multiple risk levels, adjusts the project duration and resource allocation according to the analysis results, and generates a risk impact adjustment result; The bill of quantities update module dynamically compiles and updates the bill of quantities based on the resource allocation adjustment result and the risk impact adjustment result, and generates a dynamically adjusted bill of quantities.

[0014] Compared with the prior art, the advantages and positive effects of the present invention are as follows: In the present invention, by analyzing the dependency between tasks, a task network structure with a quantifiable scoring mechanism is constructed, so that the logical association between tasks does not only stay at the static hierarchical presentation, but introduces the difference weights of time and resource dimensions, realizes dynamic modeling of critical path dependency, and provides deep support for subsequent prediction and adjustment. The tracking of task progress changes no longer relies on the static progress comparison method, but introduces the changes of the predecessor tasks into the neural network model through the deep extraction of the actual construction data, realizes the timeliness prediction of the subsequent task changes, and improves the accuracy of delay prediction. In the process of adjusting the construction period, by constructing the optimization objective function under multiple constraints, the resource and time allocation are guided to converge in the optimal direction, which significantly improves the adjustment efficiency and decision-making scientificity under complex conditions. For uncontrollable environmental variables, such as external factors such as weather, the risk level is evaluated by multi-factor fuzzy logic, and the priority sorting mechanism is introduced to screen and rearrange the affected tasks to ensure that the construction arrangement has higher anti-interference ability. In the dynamic update process, the automatic adjustment of the bill of quantities is promoted by the structured parameter system, and the effective synchronization of the quantity data is realized through the linkage change of the task-level construction period and resources, and the timeliness and accuracy of the list data are improved. The above actions work together in the prediction, response, evaluation and execution links during the task advancement process, forming a complete closed loop from perception to adjustment, significantly improving the flexibility of the bill of quantities in response to external changes, the rationality of construction resource allocation and the dynamic adaptability of the overall engineering plan. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 It is a schematic diagram of the workflow of the present invention; Figure 2 is a flow chart of step S1 of the present invention; Figure 3 is a flow chart of step S2 of the present invention; Figure 4 is a flow chart of step S3 of the present invention; Figure 5 is a flow chart of step S4 of the present invention; Figure 6 This is a flow chart of step S5 of the present invention. DETAILED DESCRIPTION

[0016] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0017] In the description of the present invention, it should be understood that the orientation or positional relationship indicated by the terms "length", "width", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc. is based on the orientation or positional relationship shown in the drawings, and is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the present invention. In addition, in the description of the present invention, the meaning of "a plurality of" is two or more, unless otherwise specifically defined.

[0018] Please refer to Figure 1 , the present invention provides a technical solution: an AI-based automatic preparation method for a bill of quantities, including the following steps: S1: Obtain engineering construction task information, including a previous first task and a subsequent second task that depends on the previous first task. Analyze the dependency relationship between tasks according to the construction period of each first task and second task. Establish a task dependency network based on the dependency relationship and draw a dependency relationship diagram between tasks, and integrate and generate a task dependency model; S2: Based on the task dependency model, real-time track the changes of the first task and the second task in the construction progress. Take the progress change of the first task as the input of the deep neural network, predict the start time and construction period change of the second task, automatically quantify the impact of project delays, and generate a progress change analysis result; S3: Obtain the impact of the delay of the first task on the start time of the second task in the progress change analysis result. Use the constraint optimization algorithm to set the conditions of engineering task time constraints, resource constraints, and dependency relationship constraints. Analyze the optimal adjustment time of the construction period of the second task and the resource allocation method under the condition of meeting all constraint conditions. Readjust the construction period and resource allocation of the second task according to the analysis result, and generate a resource allocation adjustment result; S4: By setting fuzzy evaluation rules for multiple risk levels, evaluate the weather conditions according to the obtained current weather data, including precipitation, wind speed, and temperature. Analyze the potential risks of weather factors on construction tasks. Adjust the construction period and resource allocation according to the analysis result, and prioritize the tasks affected by the weather to generate a risk impact adjustment result; S5: Based on the resource allocation adjustment result and the risk impact adjustment result, dynamically prepare the bill of quantities, including integrating the adjusted construction period, resource allocation, and task priorities, and updating the list in real time to generate a dynamically adjusted bill of quantities; The task dependency model includes task nodes, dependency paths, and critical paths. The results of the schedule change analysis include the predicted delay duration, task delay level, and the degree of impact on related tasks. The results of the resource allocation adjustment include the duration of the compressed construction period, the resource allocation plan, and the task adjustment priority. The results of the risk impact adjustment include the weather risk level, the affected task list, and the task priority ranking. The dynamically adjusted bill of quantities includes the adjusted task duration, the optimized resource allocation list, and the task implementation priority.

[0019] See also Figure 2 ,The specific steps for obtaining the task dependency model are: S111: Based on the engineering construction task information, the start time, duration, and task identification attributes of each task are obtained, and the preceding first task and the subsequent second task that depends on the preceding first task are extracted, the time dependency relationship between the tasks is determined, the duration data of each first task and the second task are analyzed, and the task timing dependency analysis result is generated; First, the system will obtain preliminary data of the project construction, for example, the start time of the task is April 1, 2025, the duration is 10 days, and the predecessor task of the task is determined. Then, the system will analyze the duration data between the tasks and calculate the time difference between the tasks. For example, suppose the duration of the first task is 10 days, the duration of the second task is 15 days, the end date of the first task is April 11, and the start date of the second task is April 12, and the system will evaluate the dependency between these dates. Through these data analysis, the system will calculate the start time of the second task based on the completion date of the first task. If the first task is delayed, the start time of the second task will be postponed accordingly. For example, if the first task is delayed by 2 days, the start time of the second task will also be postponed by 2 days, and the start date originally scheduled for April 12 will be postponed to April 14. Finally, the system will update the start time of the second task and clarify the dependency between the tasks.

[0020] S112: According to the task timing dependency analysis results, establish a task node set and an edge set of its dependency paths, draw a dependency graph between tasks, determine whether the dependency between tasks meets the timing requirements, analyze whether there are loops between task nodes, screen valid data sets, and generate task dependency structure graph data; According to the time dependencies of tasks, start constructing a task dependency network. First, consider each task as a node and connect the time dependencies between tasks as edges to form a task network. Suppose the start time of Task 1 is April 1, 2025, and the start time of Task 2 is April 12, 2025. The dependency relationship indicates that Task 1 affects Task 2, forming an edge connection. Then, the system will analyze the task dependency graph to check for the existence of a loop structure. A loop means there is a circular dependency between tasks. For example, Task 2 depends on Task 1, and Task 1 depends on Task 2, which can lead to deadlocks. In this case, the system will detect the loop and eliminate the invalid paths, screening out the valid dependency paths that meet the timing requirements. Finally, generate task dependency structure map data to show each task node and its dependency relationship with other task nodes, forming a clear dependency network between tasks for subsequent task scheduling and progress control.

[0021] S113: Based on the task dependency structure map data, extract the sequence of task nodes and their dependency paths, and use the formula: ; Calculate the dependency score of the first task node on the second task node , integrate the time dependencies of tasks and the dependency paths between tasks to establish a task dependency model; Where, and represent the start times of the second task node and the first task node respectively, is the time difference between the two, and are the maximum time and minimum time of all tasks in the system respectively; and represent the resource consumption (such as working hours) of the second task node and the first task node respectively, is the difference in resource consumption between the two, and are the maximum resource consumption and minimum resource consumption of all tasks respectively, is a weighting coefficient representing the impact degree of resource consumption on the dependency score.

[0022] First, extract all task nodes and their related dependency paths from the task dependency structure map and arrange each node in chronological order. For example, if the start time of the first task node is April 1, 2025, and the start time of the second task node is April 12, 2025, the system will arrange them in this chronological order to ensure that the dependency relationship conforms to the actual progress. Next, the system will calculate the dependency score between these two task nodes based on factors such as the time difference and resource consumption between the first task node and the second task node.

[0023] Hypothesis: The end time of the first task node , the planned start time of the second task node , the maximum time of all tasks in the system , the minimum time of all tasks in the system , the resource consumption of the first task node hours, the resource consumption of the second task node hours, the maximum resource consumption of all tasks in the system hours, the minimum resource consumption of all tasks in the system hours, the weighting factor .

[0024] According to the formula: ; Convert the date to days and calculate: time difference , time interval .

[0025] ; Next, calculate the resource difference: resource difference , resource interval .

[0026] ; Finally, the dependency score is: ; The finally obtained dependency score . To define the strength of the dependency score, the interval and threshold of the dependency score can be defined. For example, a dependency score greater than 0.1 indicates a strong dependency relationship between tasks, while a score less than 0.1 indicates a weak dependency relationship. Based on this judgment, the dependency between the first task node and the second task node is weak, but still has a certain impact. This indicates that although there is a dependency between tasks, the difference in their start times and resource consumption is small, so the impact on the overall task progress is small. The software tools used by the system construct a task dependency model. Commonly used software tools include Microsoft Project, Primavera P6, and custom task scheduling tools. Through these tools, the system can visualize the dependency relationships between tasks and generate a dependency relationship diagram of tasks. In the diagram, each task node is connected by a directed edge, indicating the dependency relationship between tasks. The connection lines between task nodes with strong dependencies are thicker, indicating strong dependencies; the connection lines between task nodes with weak dependencies are thinner. Through such a dependency network, the project manager can clearly see the time arrangements of each task node and their mutual influences, and thus effectively perform task scheduling and resource allocation.

[0027] Please refer to Figure 3 , and the steps for obtaining the progress change analysis result are specifically as follows: S211: Obtain the construction progress data of the first task node and the second task node, including the actual duration and progress deviation. Based on the task dependency model, integrate the progress change of the first task node as the pre-input data of the deep neural network to generate the progress change input data; First, input the progress change data of the first task node into the deep neural network model. In actual operation, the progress change of the task node is usually based on the deviation between the actual work progress and the predetermined plan. For example, assume that the planned duration of the first task node is 10 days and the actual completed duration is 12 days. Then the deviation between the actual duration and the planned duration is 2 days, and this deviation will be used as the model input. The deep neural network will predict the impact on subsequent tasks based on this information. Next, the data preprocessing steps ensure the accuracy of all input data, which means that all duration deviations, time differences, and progress changes need to be calculated according to the actual situation. In this process, by cleaning the historical data of the task, ensure that there are no outliers, and normalize the data so that all input data have the same scale and unit. After data cleaning and standardization, ensure that the format of the input data is unified, which is convenient for the subsequent model to perform effective training and prediction. Finally, the generated progress change input data can completely describe the dynamic situation of the progress change of the first task node.

[0028] S212: Input the progress change input data into the deep neural network for training, predict the start time and duration change of the second task node, calculate the new start time and duration of the second task node based on the progress change of the first task node, and output the task progress prediction result; The deep neural network model is trained with input data of progress changes. The training process includes multiple iterations, and the parameters (weights and biases) in the model are optimized by minimizing the error function. Specifically, the training process of the deep neural network starts with the progress change information extracted from the task historical data, and performs backpropagation on each layer of the network, adjusting the weights and biases layer by layer to minimize the prediction error of the network. Suppose the progress of the first task node is delayed by 2 days. Each layer in the network will learn based on this data. The model processes the data through the neuron activation function of each layer and adjusts the network parameters through weight updates. This process enables the neural network to gradually recognize the temporal dependencies between tasks, that is, how the progress change of the first task node affects the subsequent task nodes. For example, during the training process, the deep neural network will take the delay time of the first task node as an input feature, and at the same time combine the start time and duration of the second task node in the historical data. The model will learn the specific impact of the change of the first task on the start time and duration of the second task. After each training, the model adjusts the network parameters according to the output error to optimize the prediction accuracy. Through multiple iterations, the model can capture the non-linear impact of the progress change of the first task node on the progress of the second task node, and finally optimize the prediction accuracy to a sufficiently good level through backpropagation. When the model training is completed, the input data of progress changes will be used as the input of the model. Based on the progress change of the first task node, the model predicts the changes in the start time and duration of the second task node. Through the output layer of the neural network, the system will generate the new start time and duration adjustment values of the second task node. This process not only depends on the input delay time data, but also takes into account the complex dependencies between task nodes, so as to generate accurate task progress prediction results.

[0029] S213: According to the task progress prediction result, quantify the impact of the change in the duration of the second task node on the overall project progress, including the delay time and delay amplitude, and generate a progress change analysis result; Quantitatively analyze the change in the duration of the second task node. If the start time of the second task is postponed, affecting the subsequent progress of the overall project, the system will calculate the delay time according to the prediction result. For example, assume that the start time of the second task is postponed by 2 days and affects the subsequent 3 tasks. The system will quantify the overall project delay amplitude by analyzing the duration changes of each subsequent task. Assume that the duration of the subsequent tasks is extended by 3 days. Then the overall project delay time is the 2-day postponed start time plus the duration delays of the 3 tasks, totaling 11 days. In this way, the system can quantify the specific impact of the project delay on the overall project progress and provide data support for subsequent adjustments.

[0030] Please refer to Figure 4 , and the steps for obtaining the resource allocation adjustment result are specifically as follows: S311: Based on the progress change analysis result, collect and organize the delay data of the first task node, including the deviation between the completion time and the planned completion time, analyze the impact of the delay of the first task on the start time of the second task node, and generate the delay impact data of the first task; Collect and organize the delay data of the first task node, including the deviation between the actual completion time and the planned completion time, analyze the impact of the delay of the first task on the start time of the second task node, and generate the delay impact data of the first task. In this sub-step, first collect the actual completion time and the planned completion time of the first task node and organize them as delay data. Assuming that the planned completion time of the first task node is April 10, 2025, the actual completion time is April 12, 2025, and the delay time is 2 days, this deviation will be used as input data affecting the start time of the second task. Next, by calculating the impact of the delay time on the second task node, for example, if the start time of the second task was originally scheduled for April 12, 2025, the deep neural network model will infer the change in the start time of the second task through its dependency and delay data. At this point, the system will analyze the specific impact of the delay of the first task on the second task, especially considering factors such as interdependence between tasks and resource allocation.

[0031] S312: Based on the delay impact data of the first task, the constraint optimization algorithm is used to set the conditions of engineering task time constraints, resource constraints and dependency constraints, and the construction period and resource allocation of the second task node are constrained. According to the set conditions, the formula is used: ; Calculate the objective function of the second task duration under all constraints , determine the optimal adjustment time according to the objective function and generate the optimal duration and resource allocation list; in, is the adjusted duration of the second task The values ​​are normalized within the allowed time range. is the adjusted resource requirement of the second task The value normalized within the range of allocatable resources, is the dependency score of the first task on the second task The normalized value, , , : They are the weight parameters of duration, resources and dependency respectively. is the weight parameter of the construction period. It represents the importance of the project management to the construction period. If a project is in a rush stage or must be delivered on time, Set relatively high to preferentially shorten the time required for tasks. Conversely, if the project time is relatively loose, or the construction period is not the main source of pressure, then it can be set relatively low. is the weight parameter of resources. It reflects the degree of attention of the management side to resource savings or utilization efficiency. If project resources are tight, such as insufficient manpower, limited equipment, or there are strict cost control objectives, should be set relatively high to promote the system to preferentially optimize resource usage. Relatively speaking, if resources are sufficient and the usage cost is acceptable, this parameter can be appropriately reduced. is the weight parameter of task dependence. It indicates the degree to which the system should consider the impact relationship between previous and subsequent tasks during optimization. When there are a large number of critical path tasks in the project, or there is a strong logical sequence between tasks (for example, the structure can only be built after the foundation construction is completed), then should be set relatively high to emphasize maintaining a reasonable task sequence and construction process. If the dependence between tasks is weak or they can be processed in parallel, this value can be reduced.

[0032] All parameters need to be normalized so that they are in the range of 0 to 1: Normalization of time constraints: , Normalization of resource constraints: , Normalization of dependence relationships: ; Time constraint conditions: , Resource constraint conditions: , Dependence relationship constraints: .

[0033] Among them, is the maximum allowable construction period of the task time constraint, is the minimum allowable construction period of the task time constraint, is the maximum available resource of the task resource constraint, is the minimum available resource of the task resource constraint, is the dependence score of the first task node on the second task node.

[0034] During the optimization process, the goal is to adjust and to minimize the objective function , while satisfying all constraint conditions. Through calculation, the optimal construction period adjustment time and resource allocation plan are obtained, and a reasonable adjustment plan is obtained under the constraint conditions.

[0035] Assume: , , , , , , , , , , .

[0036] Among them, is the construction period of the first task, is the delay time of the first task, is the original construction period of the second task, is the resource requirement of the second task.

[0037] Step 1: Normalize time ; Step 2: Normalize resources ; Step 3: Normalize the dependency relationship ; Step 4: Calculate the objective function .

[0038] In order to obtain the optimal construction period adjustment time, round the result of the objective function calculation to ensure that the construction period is an integer value. After adjustment, the optimal adjustment time for the construction period of the second task is . Therefore, the final adjusted construction period of the second task is 53 days, and the resource allocation plan is also adjusted accordingly to meet the constraints of time, resources, and dependency relationships.

[0039] S313: Based on the optimal construction period and resource allocation list, readjust the construction period and resource allocation of the second task to generate a resource allocation adjustment result; Adjust the specific construction period and resource allocation of the second task. For example, assume that the original construction period of the second task is 10 days, and according to the optimization result, the construction period needs to be extended to 13 days. Also, to meet the resource constraints, the system may also adjust the required equipment and personnel configuration. Specifically, assume that the second task requires an additional 2 pieces of equipment and 4 workers to complete within the new construction period. The system will adjust the resource allocation to meet this requirement and ensure that the task can be completed within the new construction period. Finally, the system generates a resource allocation adjustment result based on the adjusted construction period and resource allocation, which can provide the project manager with an optimized task scheduling plan to help reasonably arrange the execution of subsequent tasks and resource allocation.

[0040] Please refer to Figure 5 , the specific steps for obtaining the risk impact adjustment result are as follows: S411: Obtain the current weather data, including precipitation, wind speed, and temperature, combine it with the progress and resource requirement information of the construction task to generate a weather dataset; Obtain the current weather data, including precipitation, wind speed, and temperature. For this process, it is necessary to first connect to the corresponding automatic meteorological acquisition device at the construction site and perform periodic sampling processing on the weather factors in the area. The sampling time interval is set to 1 hour, and the sample period is 7 consecutive days. Format the precipitation in millimeters, the wind speed in meters per second, and the temperature in degrees Celsius and store them in the meteorological data structure. At the same time, obtain the current task status data from the construction task scheduling system, including task number, task type, current progress percentage, remaining construction period, details of required resources, etc., and perform timestamp matching with the meteorological data collected at that time to form combined data items. Further obtain the task resource configuration data, including actual parameters such as the number of operating personnel, equipment usage hours, and volume of required materials. Construct a data representation by merging fields for the above three types of data to form a three-dimensional structure array, where the task is the primary key, time is the secondary key, and parameters are the sub-keys. This structure represents the weather environment, task status, and resource demand status of any task at any time. For example, at 9 am on August 15, 2024, task A107 at a certain construction site is in the main construction stage with a progress of 62%, corresponding to a precipitation of 14 mm, a wind speed of 9 m / s, a temperature of 29°C, 10 operating personnel, and 4 resource devices. This data, combined with similar data at 8:00 and 10:00, can form a sequence data set with an hourly step size, which is used as the basic input for subsequent risk calculations, adjustment calculations, etc., and finally forms a weather data set.

[0041] S412: Based on the weather data set, according to the fuzzy evaluation rules set for multiple risk levels, match the precipitation, wind speed, and temperature weather factors with the risk levels, using the formula: ; Calculate the fuzzy risk value affected by the weather , analyze the potential construction risks of each weather condition, and evaluate the impact degree of the weather condition on the construction task to generate a weather risk assessment result; Among them, represents the actual precipitation, represents the actual wind speed, represents the actual temperature, , , respectively represent the intermediate control values of the corresponding weather factors, , , respectively represent the maximum reference values of each factor in the set risk levels, with the units of millimeters, meters per second, and degrees Celsius respectively. In the formula, through the sum of squared absolute deviations and the normalization process for the maximum value, a concentrated reflection of the risk degree is formed.

[0042] According to the fuzzy assessment rules for the set multiple risk levels, the precipitation, wind speed, and temperature weather factors will be matched with the risk level, and the fuzzy risk value will be calculated using the following formula: ; in, It represents the measured precipitation in millimeters, which is measured hourly by the weather station; is the reference control value, set to 25 mm, which is derived from the recommended value of the national construction site rainy day construction specification; is the maximum annual precipitation in the region, set to 50 mm, derived from the monthly extreme value statistics released by the local meteorological bureau; is the wind speed in m / s, measured by an on-site anemometer, and currently set to 13m / s; The control value is 15m / s, The maximum reference value is 30m / s; Indicates the current temperature, which is 30°C, collected by automatic temperature sensing equipment; The construction temperature control value is 25°C. The maximum safe operating temperature is 40°C. Substituting the measured data at noon on August 15, 2024, we get: ; This value is in the "medium-high risk" section of the risk level classification standard, ranging from 1.2 to 2.0. The risk classification standard is based on the research data on the impact of relevant meteorological parameters on construction in the "Guidelines for Meteorological Safety Operations in Construction Engineering". The risk levels are defined as: low risk 0-0.5, medium risk 0.5-1.2, medium-high risk 1.2-2.0, and high risk above 2.0. The medium-high risk values ​​will be used as a reference for the construction period and resource allocation adjustment strategy to obtain the weather risk assessment results.

[0043] By converting the deviation values ​​of precipitation, wind speed and temperature into a sum of squares and then normalizing them with the maximum reference value, we can effectively integrate the dimensions and fluctuation ranges of different meteorological factors, making the results comparable and scalable between different seasons and regions. Compared with traditional weighted average or linear superposition, this method can better express the intensity of risk aggregation when multiple factors are linked, and facilitates the formulation of differentiated task response strategies.

[0044] S413: According to the weather risk assessment results, adjust the construction period and resource allocation, prioritize the tasks according to the degree to which they are affected by the weather, and generate risk impact adjustment results; First, call the corresponding risk value item by item according to the task number, and determine which risk level interval it is in. If it is a task with medium risk or above, it will enter the adjustment list. The first step of the adjustment process is to extend the original remaining duration of the task. The extension amplitude is set according to the risk level. Medium-high risk corresponds to 30% of the original remaining duration. For example, if the remaining duration of task A107 is 3 days, the number of extended days is 3×0.3 = 0.9 days, which is rounded up to 1 day. The second step is to determine whether resource allocation can be increased according to the required resource type of the task. The resource allocation strategy is 20% of the original configuration. The original configured number of workers for task A107 is 12, and the number of construction machinery is 5. After the increase, they are 14 and 6 respectively. The third step is to re-evaluate the task priority. Establish a priority scoring rule based on the exposure degree of the construction area, task category, equipment transfer difficulty, and personnel coordination complexity. On this basis, adjust A107 to the preposed priority position to participate in the sorting of the scheduling system. After re-sorting, export the task adjustment list for updating the construction plan, and finally obtain the risk impact adjustment result. This result shows that when the risk value R is in the medium-high level, it will directly trigger the extension and resource allocation increase behaviors. The higher the value, the greater the corresponding changes in the construction period and resources. It plays a decisive role in the construction progress control. The calculated numerical result has a direct operational meaning and can be used to guide the on-site schedulers to update the construction strategy immediately.

[0045] Please refer to Figure 6 , and the specific steps for obtaining the dynamically adjusted bill of quantities are as follows: S511: Based on the resource allocation adjustment result and the risk impact adjustment result, call the construction period adjustment amount, resource configuration change value, and priority order in each task node, perform unified numbering association on the corresponding task entries, and generate a task adjustment parameter structure; First, traverse the task number list. During the traversal process, match the corresponding adjustment records. For example, the original planned construction period of task number T023 is 8 days, which is extended by 2 days after adjustment. The resource usage changes from 4 devices to 5 devices, and the priority changes from 5 to 3 at the same time. This data is recorded in the structure as T023: construction period +2, device +1, priority 3. All the matched records are encapsulated into the task structure dictionary and marked with a timestamp to distinguish versions. Subsequently, compare this structure with the entry index of the bill of quantities, locate the position with the original task number in the list as the primary key, and bind the three adjustment values to the additional data segment of the record where the task number is located. Then, normalize the data format of the fields in the generated structure to confirm the unit standard of the values (such as the construction period is in days, the resources are in devices, and the priority is represented by natural numbers). Finally, rearrange the data structure according to the generation order to obtain the task adjustment parameter structure.

[0046] S512: Adjust the parameter structure according to the tasks, match the duration value, resource data, and priority order of each task to the field positions associated with the corresponding task numbers in the bill of quantities in sequence, mark the corresponding fields in the original bill of quantities as updatable according to the field positions, and generate list field update data; First, retrieve the corresponding entry row in the bill of quantities through the task number. After successful retrieval and positioning, perform a field-level comparison between the three adjusted data and the original list fields and record the differences. For example, if the "duration field" in the original list of a certain task is 8 days and is 10 days in the adjusted structure, the system records that this field is "marked for update". This marking information is synchronously recorded in the list cell data management system for subsequent list update operations. At the same time, write the change source type in each field change record. The change type of the "duration field" is set to "source: risk impact adjustment", the change type of the resource field is set to "source: resource allocation adjustment", and the priority change record is "source: two-way intersection". The change value of each piece of data in the above fields is attached to the end of the record body in the form of "original value - present value". For example, the duration field change is "8 → 10". After completing the record operation, construct a complete list change data set as the basis for subsequent updates to obtain list field update data.

[0047] S513: Invoke the list field update data, perform a data replacement operation on the fields involved in the update in the original bill of quantities structure, and reassign the list version structure after the replacement. The number is represented by the current date plus the version number, covering the display fields of the original list file to generate a dynamically adjusted bill of quantities; First, read the field change entries in the update data set, locate the corresponding record row in the original list according to the task number, find the field name marked as "to be updated" in each row, and confirm whether its original value is consistent with the updated value. If not, directly overwrite the original value with the updated value and update the modification timestamp field of the corresponding field to record the time of this data change. For example, the "duration field" of task T023 is updated to 10 days, and the field status is changed to "updated". After the overwrite is completed, perform a number reorganization on the entire list. Use the current date (such as 20250403) plus a three-digit serial number to form a new version number "20250403 - 001", and save this version to the database for subsequent calls. At the same time, keep the list hierarchy structure unchanged, reorganize the task numbers, project locations, and resource item order to ensure the integrity of the data logical structure, and finally output and save the processed list content to generate a dynamically adjusted bill of quantities.

[0048] An AI-based automatic bill of quantities compilation system. The AI-based automatic bill of quantities compilation system is used to execute the above-mentioned AI-based automatic bill of quantities compilation method. The system includes: The task dependency modeling module obtains the engineering construction task information, analyzes the dependency relationships between construction tasks, and constructs a task dependency model based on the dependency relationships; The progress prediction and change analysis module, based on the task dependency model, tracks the task changes in the construction progress in real time, uses a neural network to predict task changes, and generates a progress change analysis result; The task duration optimization module obtains the mutual influence of task delays in the progress change analysis result, analyzes the optimal adjustment time of the task duration under the satisfaction of the constraint conditions by setting the constraint conditions of the engineering tasks, and makes corresponding adjustments to generate a resource allocation adjustment result; The risk assessment and impact adjustment module analyzes the potential risks of weather factors on construction tasks by setting fuzzy assessment rules for various risk levels, adjusts the project duration and resource allocation according to the analysis results, and generates a risk impact adjustment result; The bill of quantities update module dynamically compiles and updates the bill of quantities based on the resource allocation adjustment result and the risk impact adjustment result, and generates a dynamically adjusted bill of quantities.

[0049] The above is only a preferred embodiment of the present invention, and does not limit the present invention in other forms. Any person skilled in the art may use the technical content disclosed above to make changes or modifications into equivalent embodiments with equivalent changes and apply them to other fields. However, as long as it does not depart from the technical solution content of the present invention, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention still fall within the protection scope of the technical solution of the present invention.

Claims

1. An AI-based automatic preparation method for a bill of quantities, characterized in that, It includes the following steps: S1: Obtain the engineering construction task information, analyze the dependency relationships between construction tasks, and construct a task dependency model based on the dependency relationships; S2: Based on the task dependency model, track the task changes in the construction progress in real time, use a neural network to predict task changes, and generate a progress change analysis result; S3: Obtain the mutual influence of task delays in the progress change analysis result, analyze the optimal adjustment time of the task duration under the condition of meeting the constraint conditions by setting the constraint conditions of engineering tasks, and make corresponding adjustments to generate a resource allocation adjustment result; S4: Analyze the potential risks of weather factors on construction tasks by setting fuzzy evaluation rules for multiple risk levels, adjust the construction period and resource allocation according to the analysis results, and generate a risk impact adjustment result; S5: Based on the resource allocation adjustment result and the risk impact adjustment result, dynamically compile and update the bill of quantities to generate a dynamically adjusted bill of quantities.

2. The AI-based automatic preparation method for the bill of quantities according to claim 1, wherein The task dependency model includes task nodes, dependency paths, and critical paths. The progress change analysis result includes the predicted delay duration, task delay level, and the degree of influence on associated tasks. The resource allocation adjustment result includes the duration compression duration, resource allocation plan, and task adjustment priority. The risk impact adjustment result includes the weather risk level, the list of affected tasks, and the task priority ranking. The dynamically adjusted bill of quantities includes the adjusted task duration, the optimized resource allocation list, and the task implementation priority order.

3. The AI-based automatic preparation method for the bill of quantities according to claim 1, characterized in that, The specific steps for obtaining the task dependency model are as follows: S111: Based on the engineering construction task information, obtain the start time, duration, and task identification attributes of each task, extract the preceding first task and the subsequent second task that depends on the preceding first task, judge the time dependency relationship between tasks, and analyze the duration data of each first task and the second task to generate a task time sequence dependency analysis result; S112: According to the task time sequence dependency analysis result, establish a set of task nodes and an edge set of dependency paths, draw a dependency relationship diagram between tasks, judge whether the dependency relationship between tasks meets the time sequence requirements, analyze whether there are loops between task nodes, and screen the data set to generate task dependency structure map data; S113: Based on the task dependency structure map data, extract the sequence of task nodes. By calculating the time difference between the first task node and the second task node, and combining the maximum resource consumption and minimum resource consumption of the first task and the second task, quantitatively evaluate the dependency relationship between the two task nodes, and establish a task dependency model.

4. The AI-based automatic preparation method of the bill of quantities according to claim 3, characterized in that, The specific steps for obtaining the progress change analysis result are as follows: S211: Obtain the construction progress data of the first task node and the second task node, including the duration and progress deviation. Based on the task dependency model, integrate the progress change of the first task node as the pre-input data of the deep neural network to generate progress change input data; S212: Input the progress change input data into a deep neural network for training to predict the start time and duration change of the second task node, calculate the new start time and duration of the second task node based on the progress change of the first task node, and output the task progress prediction result; S213: According to the task progress prediction result, quantify the impact of the duration change of the second task node on the overall project progress, including the delay time and delay amplitude, and generate a progress change analysis result.

5. The AI-based automatic preparation method for the bill of quantities according to claim 4, characterized in that, The steps for obtaining the resource allocation adjustment result are specifically as follows: S311: Based on the progress change analysis result, collect and organize the delay data of the first task node, including the deviation between the completion time and the planned completion time, analyze the impact of the delay of the first task on the start time of the second task node, and generate the first task delay impact data; S312: Based on the first task delay impact data, use a constraint optimization algorithm to set the conditions of engineering task time constraints, resource constraints, and dependency constraints, and perform constraint analysis on the duration and resource allocation of the second task node. According to the set conditions, use the formula: ； Objective function for calculating the second task duration under all constraints , determine the optimal adjustment time according to the objective function, and generate a list of optimal duration and resource allocation; Among them, is the duration after the second task is adjusted the value after normalization within the allowed time range, is the resource requirement after the second task is adjusted the value after normalization within the allocable resource range, is the dependency score of the first task on the second task the value after normalization, , , : are the weight parameters for duration, resources, and dependency respectively; S313: Based on the optimal duration and resource allocation list, readjust the duration and resource allocation of the second task to generate a resource allocation adjustment result.

6. The AI-based automatic preparation method of the bill of quantities according to claim 5, wherein, The steps for obtaining the risk impact adjustment result are specifically as follows: S411: Obtain the current weather data, including precipitation, wind speed, and temperature, combine it with the progress and resource requirement information of the construction task to generate a weather data set; S412: Based on the weather data set, according to the fuzzy evaluation rules of multiple risk levels set, match the weather factors of precipitation, wind speed, and temperature with the risk levels, and use the formula: ； Calculate the fuzzy risk value of weather impact , analyze the potential construction risks of each weather condition, evaluate the impact degree of the weather condition on the construction tasks, and generate the weather risk assessment results; Among them, represents precipitation, represents wind speed, represents temperature, 、 、 respectively represent the intermediate control values of the corresponding weather factors, 、 、 respectively represent the maximum reference values of each factor in the set risk level; S413: According to the weather risk assessment result, adjust the duration and resource allocation, prioritize the tasks according to the degree of influence of the tasks by the weather, and generate a risk impact adjustment result.

7. The AI-based automatic preparation method for bill of quantities according to claim 6, wherein The steps for obtaining the dynamically adjusted bill of quantities are specifically as follows: S511: Based on the resource allocation adjustment result and the risk impact adjustment result, call the duration adjustment amount, resource configuration change value, and priority order in each task node, perform unified numbering association on the corresponding task items to generate a task adjustment parameter structure; S512: According to the task adjustment parameter structure, match the duration value, resource data, and priority order of each task to the field positions associated with the corresponding task numbers in the bill of quantities in sequence, and mark the corresponding fields in the original bill of quantities as updatable according to the field positions to generate list field update data; S513: Call the list field update data, perform data replacement operations on the fields that need to be updated in the original bill of quantities structure, and reassign the list version structure after the replacement is completed. The number is represented by the current date plus the version number, covering the display fields of the original list file, and generate a dynamically adjusted bill of quantities.

8. An AI-based automatic bill of quantities preparation system, characterized in that, According to the AI-based automatic bill of quantities compilation method according to any one of claims 1-7, the system includes: The task dependency modeling module obtains the engineering construction task information, analyzes the dependency relationships between construction tasks, and constructs a task dependency model based on the dependency relationships; The schedule prediction and change analysis module, based on the task dependency model, real-time tracks the task changes in the construction schedule, uses neural networks to predict task changes, and generates a schedule change analysis result; The task duration optimization module obtains the mutual influence of task delays in the schedule change analysis result, analyzes the optimal adjustment time of the task duration under the satisfaction of the constraint conditions by setting the constraint conditions of the engineering tasks, and makes corresponding adjustments to generate a resource allocation adjustment result; The risk assessment and impact adjustment module analyzes the potential risks of weather factors on construction tasks by setting fuzzy assessment rules for various risk levels, adjusts the project duration and resource allocation according to the analysis results, and generates a risk impact adjustment result; The bill of quantities update module dynamically compiles and updates the bill of quantities based on the resource allocation adjustment result and the risk impact adjustment result, and generates a dynamically adjusted bill of quantities.

Images