Intelligent engineering cost dynamic calculation and cost control method and system

Abstract

The invention relates to the technical field of engineering cost dynamic measurement and calculation, in particular to an intelligent engineering cost dynamic measurement and calculation and cost control method and system. Comprehensive influence factors of nodes can be extracted from a complex construction task incidence relation through a graph convolutional network; comprising resource intensity, progress sequence and potential resource conflict problems, so that accurate identification of key construction paths is improved, the real-time feedback capability of project progress and cost is enhanced, dependence of a traditional budget model on static factors is effectively avoided, project decision is optimized according to real-time changes, and the construction efficiency is improved. The efficient distribution of budget and resources is ensured, through the application of the non-dominated sorting genetic algorithm, and according to the residual budget, the construction period and the resource value, comprehensive evaluation and node screening optimization are carried out, so that project management can flexibly adjust the budget and the construction period based on the actual situation, the situation of unreasonable fund scheduling or resource waste is avoided, and the project management efficiency is improved. And the accuracy and flexibility of engineering project management are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of dynamic calculation of engineering costs, and in particular to an intelligent dynamic calculation and cost control method and system for engineering costs. Background Art

[0002] The technical field of dynamic measurement of construction cost aims to update the construction cost estimate in real time and perform measurement and control based on multiple factors such as progress changes, resource utilization, and design changes during the entire implementation process of construction projects. It involves budget preparation, cost accounting, change management, dynamic forecasting and control measures, aiming to achieve dynamic management and reasonable control of project costs, and avoid budget overruns, waste, and unreasonable fund allocation.

[0003] An intelligent dynamic cost estimation and cost control method for engineering projects aims to establish a dynamic cost estimation system that can automatically respond to factors such as design changes, schedule adjustments, material price fluctuations, and changes in resource inputs, and to achieve immediate correction and precise control of the total project cost, thereby achieving the effects of rational use of engineering expenses, timely and accurate cost accounting, and efficient execution of budget control, thereby achieving optimal allocation of engineering investment and effective prevention of risks.

[0004] In existing technologies, dynamic measurement and cost control of project costs often rely on traditional budget preparation and regular update mechanisms, which are slow to respond to various changes in the project and lack the ability to effectively process multi-dimensional, real-time data. As a result, it is impossible to update the budget and schedule in a timely and accurate manner, and it is difficult to reflect the actual situation in project implementation. In addition, existing technologies usually focus on static budget accounting and cannot dynamically adjust plans according to the progress of the project. Serious problems such as budget overruns or construction delays are often discovered in the middle of the project. It is impossible to effectively prevent the problem of unreasonable fund scheduling, which affects the overall economic benefits and execution efficiency of the project. Summary of the Invention

[0005] The purpose of the present invention is to solve the shortcomings of the prior art and to propose an intelligent engineering cost dynamic calculation and cost control method and system.

[0006] In order to achieve the above object, the present invention adopts the following technical solution: an intelligent engineering cost dynamic calculation and cost control method, comprising the following steps:

[0007] S1: Based on the construction task information, the project quantity, unit price, construction period and resource requirements of the foundation, main structure and decoration units are extracted, the node cost is calculated, the node resource density is generated through proportional coefficient calculation, the construction sequence and resource conflict relationship are combined, and a construction task association diagram is generated;

[0008] S2: Based on the construction task association graph, a graph convolutional network is used to extract node costs, resource intensity, sequence relationships, and resource conflicts to generate a comprehensive node impact matrix. An edge weight matrix is generated through dependency judgment. Combined with the schedule synchronization table and the cost prediction unit, feature weighted accumulation and sorting are performed to obtain a comprehensive construction critical path evaluation matrix.

[0009] S3: Based on the comprehensive evaluation matrix of the construction critical path, the actual progress, expenditure, and resource usage of the task are extracted, the node progress difference, expenditure deviation, and resource excess rate are calculated, abnormal nodes are screened, the dependent edge set is extracted, and a dynamic abnormal node association subgraph is generated;

[0010] S4: Based on the dynamic abnormal node association subgraph, a non-dominated sorting genetic algorithm is used to extract the remaining budget, duration, and resource values, perform budget and duration judgment, screen optimization nodes, and generate a budget and duration joint optimization adjustment plan by combining resource allocation and remaining resource accounting;

[0011] S5: Based on the budget-construction-period joint optimization and adjustment plan, extract the remaining budget, construction period and resource change rate, calculate the budget proportion and construction period tolerance, screen the correction combination, combine the cost adjustment path, and generate the project dynamic budget and construction period optimization execution list.

[0012] As a further solution of the present invention, the specific steps of generating the construction task association relationship diagram are:

[0013] Based on the construction task information, the engineering quantity, unit price, construction period, and resource requirements of the foundation unit, main unit, and decoration unit are extracted. The engineering quantity and unit price values are multiplied one by one to form the node budget cost. The node budget cost and node identification index are used to organize the mapping relationship and establish the node budget cost set.

[0014] Based on the node budget cost set, the engineering quantity and resource requirement quantity corresponding to each node are extracted. The node engineering quantity is divided by the resource requirement quantity to perform a ratio calculation. The ratio calculation value is then associated with the node budget cost to form a combined data set, generating a node resource intensity set.

[0015] Based on the node resource density set, the construction sequence and resource conflict relationship between nodes are extracted, the dependency relationship is classified using sequence numbering, and the conflict relationship identification is formed by cross-comparison of resource occupancy quantity. The sequence dependency edge and the conflict dependency edge are connected to establish the construction task association relationship diagram.

[0016] As a further solution of the present invention, the specific steps of generating the construction critical path comprehensive evaluation matrix are:

[0017] Based on the construction task association graph, a graph convolutional network is used to extract the node budget cost, resource intensity, construction sequence relationship, and resource conflict relationship. The node budget cost and resource intensity are normalized using standard processing to form a unified node indicator value set and construct a node comprehensive indicator matrix.

[0018] Based on the node comprehensive indicator matrix, the node order relationship and conflict relationship are extracted, the node sorting number is used to match the order dependency, the resource conflict ratio is used to determine the priority of the conflicting edge, the corresponding weight record table of the node and edge is generated, and the node edge weight matrix set is established;

[0019] Based on the node-edge weight matrix set, the node stage attributes, budget usage progress, and resource consumption level are extracted, and weighted accumulation is performed using the node comprehensive weight and the stage budget completion rate. The accumulated results are used to perform descending sorting of the node priorities to generate a comprehensive evaluation matrix for the construction critical path.

[0020] As a further solution of the present invention, the graph convolutional network is based on the formula:

[0021]

[0022] Among them: H (l+1) represents the comprehensive indicator matrix of the l+1th layer node, σ represents the nonlinear activation function, represents the degree matrix of the construction task association graph containing self-loops, represents the adjacency matrix of the construction task association graph containing self-loops, α represents the weight coefficient of the node budget cost attribute, represents the normalized characteristic matrix of the budget cost standard of the l-th layer node, β represents the weight coefficient of the node resource intensity attribute, represents the standard normalized feature matrix of the resource intensity of the l-th layer node, W (l) Represents the trainable weight matrix of layer l.

[0023] As a further solution of the present invention, the specific steps of generating the dynamic abnormal node association subgraph are:

[0024] Based on the comprehensive evaluation matrix of the construction critical path, the actual progress value, actual expenditure value, and actual resource usage value of the tasks corresponding to each node are extracted. The progress offset is calculated by subtracting the actual progress from the node's planned progress, the expenditure offset is calculated by subtracting the actual expenditure from the node's budgeted expenditure, and the resource difference rate is calculated by subtracting the actual usage from the node's planned resources, thereby generating a node difference rate indicator set.

[0025] Based on the node difference rate indicator set, the nodes with progress offset, expenditure offset, and resource difference rate exceeding the preset value are screened, the corresponding sequential dependency edge relationships are extracted using node index matching, the sequential edges are screened and sorted and merged, and a set of abnormal node dependency edges is generated;

[0026] Based on the abnormal node dependent edge set, the abnormal nodes and dependent edges are connected, the nodes are arranged in order and numbered according to the edge connection order, a bidirectional index table of the node set and the edge set is established, the subgraph connection relationship is drawn, and a dynamic abnormal node associated subgraph is established.

[0027] As a further solution of the present invention, the specific steps of generating the budget duration joint optimization adjustment plan are:

[0028] Based on the dynamic abnormal node association subgraph, a non-dominated sorting genetic algorithm is used to extract the remaining budget value, remaining duration value, and remaining resource value corresponding to the abnormal node, calculate the remaining budget by subtracting the accumulated expenditure value from the initial budget value of the node, calculate the remaining duration by subtracting the number of completed duration days from the initial duration value of the node, and calculate the remaining resource amount by subtracting the node resource plan amount from the used amount, thereby generating a remaining resource amount set for the abnormal node;

[0029] Based on the remaining resource amount set of the abnormal nodes, the nodes whose remaining budget ratio is lower than the budget threshold and the nodes whose remaining duration ratio is lower than the duration threshold are screened, the high-priority nodes are screened according to the node resource remaining ratio, the node priorities are rearranged according to the combined value of budget and duration, and a budget-time optimized node set is generated;

[0030] Based on the budget duration optimization node set, the node remaining budget, remaining duration, and remaining resource amount are extracted, and intersection screening is performed using the node budget remaining amount and the node duration remaining amount. The node combination is extracted to form a resource supplement path plan, the combination optimization data table is sorted out, and a budget duration joint optimization adjustment plan is established.

[0031] As a further solution of the present invention, the non-dominated sorting genetic algorithm is according to the formula:

[0032]

[0033] Where: f i represents the comprehensive remaining index value of the i-th abnormal node, w1 represents the importance weight coefficient of the remaining budget ratio, w2 represents the importance weight coefficient of the remaining duration ratio, w3 represents the importance weight coefficient of the remaining resource ratio, w4 represents the importance weight coefficient of the progress completion correction item, and RB i Indicates the remaining budget value of the i-th abnormal node, IB i represents the total initial budget of the i-th abnormal node, RD iIndicates the remaining duration value of the i-th abnormal node, ID i represents the initial duration of the i-th abnormal node, RR i Indicates the remaining resource value of the i-th abnormal node, IR i represents the total planned resources of the ith abnormal node, PC i Represents the progress completion percentage correction item of the i-th abnormal node.

[0034] As a further solution of the present invention, the specific steps of generating the project dynamic budget and construction period optimization execution list are as follows:

[0035] Based on the budget duration joint optimization and adjustment plan, the remaining budget value, remaining duration value, and resource change rate value of each node are extracted, the remaining budget value and the project approved budget value are used to perform a ratio calculation and record the budget proportion, the remaining duration value and the approved duration value are used to perform a ratio calculation and record the duration tolerance, and the resource change rate value and the node standard resource ratio are used to perform a deviation rate calculation to generate a budget duration ratio indicator set;

[0036] Based on the budget-to-duration ratio indicator set, a node set whose budget ratio is less than a set budget lower limit and whose duration tolerance is less than a set duration lower limit is selected, the node set is sorted in ascending order by the budget deviation rate and budget priority nodes are selected, the node set is sorted in ascending order by the duration deviation rate and duration priority nodes are selected, an intersection set of the budget priority nodes and the duration priority nodes is extracted, and a budget-to-duration correction node combination is generated;

[0037] Based on the budget duration correction node combination, the remaining budget value, remaining duration value, and resource change rate value of the intersection node are extracted, and the budget adjustment amount is generated by adding up the remaining budget values of the nodes, and the duration adjustment amount is generated by adding up the remaining duration values of the nodes. The budget supplement and duration supplement are integrated according to the cost adjustment path corresponding to the resource change rate of each node to establish a dynamic budget and duration optimization execution list for the project.

[0038] As a further solution of the present invention, the remaining budget value and the project approved budget value are calculated in proportion and the budget proportion is recorded. The remaining budget value of the current node is calculated to have a ratio with the project approved budget value. The calculated ratio is used to reflect the current budget execution status of the node, to determine whether the budget execution complies with the predetermined plan, and the budget proportion is recorded proportionally.

[0039] An intelligent engineering cost dynamic calculation and cost control system, which is used to execute the intelligent engineering cost dynamic calculation and cost control method, and includes:

[0040] Construction task data acquisition module: Based on construction task information, it extracts the engineering quantity, unit price, construction period, and resource requirements of the foundation unit, main unit, and decoration unit, calculates the cost of each node, and generates the resource density of the node through proportional coefficient calculation. It also combines the construction sequence and resource conflict relationship to generate a construction task association diagram;

[0041] Association graph construction module: Based on the construction task association graph, a graph convolutional network is used to extract node cost, resource density, sequence relationship, and resource conflict information, calculate the dependency relationship between each node, and generate dependency judgment data. This module then generates an edge weight matrix. The progress synchronization table and cost prediction unit perform weighted feature accumulation and sorting to generate a comprehensive evaluation matrix for the construction critical path.

[0042] Construction critical path assessment module: Based on the construction critical path comprehensive assessment matrix, it extracts the actual progress, expenditure, and resource usage of the task, calculates the progress difference, expenditure deviation, and resource excess rate of the node, filters out abnormal nodes, extracts the dependent edge set, and generates a dynamic abnormal node association subgraph;

[0043] Budget and duration optimization module: Based on the dynamic abnormal node association subgraph, a non-dominated sorting genetic algorithm is used to extract the remaining budget, duration, and resource values of each node, calculate the execution status of the budget and duration, select the optimized nodes, and combine resource allocation with remaining resource accounting to generate a budget and duration joint optimization adjustment plan;

[0044] Budget and schedule execution list module: Based on the budget and schedule joint optimization adjustment plan, extract the remaining budget, schedule and resource change rate, calculate the budget proportion and schedule tolerance, screen out the correction combination, combine the cost adjustment path, and generate the project dynamic budget and schedule optimization execution list.

[0045] Compared with the prior art, the advantages and positive effects of the present invention are:

[0046] 1. This invention uses a graph convolutional network to extract comprehensive influencing factors of nodes from complex construction task relationships, including resource intensity, progress sequence, and potential resource conflicts. This improves the accurate identification of construction critical paths and enhances the real-time feedback capability of project progress and cost.

[0047] 2. This invention combines an edge weight matrix generated by dependency judgment with a schedule synchronization table and a cost forecasting unit to perform feature-weighted accumulation and ranking of the progress, costs, and resource usage of each task in the construction project. This generates a comprehensive evaluation matrix for the construction critical path and dynamically adjusts the priority of construction tasks. This effectively avoids the traditional budget model's reliance on static factors, instead optimizing project decisions based on real-time changes and ensuring efficient allocation of budgets and resources.

[0048] 3. In the present invention, through the application of the non-dominated sorting genetic algorithm, abnormal situations such as budget overruns and construction delays can be identified in real time, and a comprehensive evaluation can be performed based on the remaining budget, construction period, and resource value to screen optimization nodes, so that project management can flexibly adjust the budget and construction period based on actual conditions, avoid unreasonable fund scheduling or waste of resources, and improve the accuracy and flexibility of engineering project management. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 It is a schematic diagram of the workflow of the present invention;

[0050] Figure 2 is a system flow chart of the present invention;

[0051] Figure 3 Schematic diagram of the system flow of the present invention. DETAILED DESCRIPTION

[0052] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to 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.

[0053] See also Figure 1 The present invention provides a technical solution: an intelligent engineering cost dynamic calculation and cost control method, comprising the following steps:

[0054] S1: Based on the construction task information, the project quantity, unit price, construction period and resource requirements of the foundation, main structure and decoration units are extracted, the node cost is calculated, the node resource density is generated through proportional coefficient calculation, the construction sequence and resource conflict relationship are combined, and a construction task association diagram is generated;

[0055] S2: Based on the construction task association graph, a graph convolutional network is used to extract node costs, resource intensity, sequence relationships, and resource conflicts to generate a comprehensive node impact matrix. An edge weight matrix is generated through dependency judgment. Combined with the schedule synchronization table and cost prediction unit, weighted feature accumulation and sorting are performed to obtain a comprehensive evaluation matrix for the construction critical path.

[0056] S3: Based on the comprehensive evaluation matrix of the construction critical path, the actual progress, expenditure, and resource usage of the task are extracted. The node progress difference, expenditure deviation, and resource excess rate are calculated. Abnormal nodes are screened, the dependent edge set is extracted, and a dynamic abnormal node association subgraph is generated.

[0057] S4: Based on the dynamic abnormal node association subgraph, a non-dominated sorting genetic algorithm is used to extract the remaining budget, duration, and resource values, make budget and duration judgments, screen the optimization nodes, and generate a budget and duration joint optimization adjustment plan by combining resource allocation and remaining resource accounting;

[0058] S5: Based on the budget-construction-duty joint optimization adjustment plan, extract the remaining budget, construction period and resource change rate, calculate the budget ratio and construction period tolerance, screen the correction combination, combine the cost adjustment path, and generate the project dynamic budget and construction period optimization execution list.

[0059] The specific steps to generate the construction task association diagram are:

[0060] Based on the construction task information, the engineering quantity, unit price, construction period, and resource requirements of the foundation unit, main unit, and decoration unit are extracted. The engineering quantity and unit price values are multiplied one by one to form the node budget cost. The node budget cost and node identification index are used to organize the mapping relationship and establish the node budget cost set.

[0061] Based on the node budget cost set, the engineering quantity and resource requirement quantity corresponding to each node are extracted. The node engineering quantity is divided by the resource requirement quantity to perform a ratio calculation. The ratio calculation value is then associated with the node budget cost to form a combined data set, generating a node resource intensity set.

[0062] Based on the node resource density set, the construction sequence and resource conflict relationship between nodes are extracted. The dependency relationship is classified using sequence numbers. The conflict relationship is identified by cross-comparison of resource occupancy quantity. The sequence dependency edge and the conflict dependency edge are connected to establish the construction task association relationship graph.

[0063] Based on the construction task information, a multi-field structured parsing method is used to extract the engineering quantity, unit price, construction period and resource requirements of the basic unit, main unit and decoration unit. The original task information is grouped and processed according to the construction unit. The field parsing function is called to perform numerical conversion operations on the engineering quantity field. Regular expressions are used to extract numerical units and convert them into floating-point types. The unit price field parsing uses a dictionary matching structure to identify and standardize the monetary unit. The resource requirement field is extracted into a key-value pair structure according to the resource type. Then, the item-by-item product algorithm is used to perform dot multiplication operations on the extracted engineering quantity and unit price. The engineering quantity value at the corresponding position in each construction unit is multiplied by the unit value one by one. The array broadcast mechanism is called to perform matrix-level multiplication on the batch construction unit data to form a list of construction node budget costs. Then, a dictionary structure index relationship is established between the node identifier and the budget cost. The node identifier is set as the key and the corresponding budget value as the value to construct a mapping set, and finally a node budget cost set is generated.

[0064] Based on the node budget cost set, the field vectorization parsing method is used to extract the engineering quantity and resource demand quantity corresponding to each node. The data structure of each node is traversed, and a two-dimensional array structure is used to store the quantity values of the engineering quantity and the corresponding resource demand. The dimension-by-division function operation logic is used to perform a value-by-value ratio calculation on each pair of engineering quantity value and resource demand quantity. The broadcast mechanism is called to expand the array dimension to ensure the consistency of the numerator and denominator alignment structure. The ratio output result is set to floating point format and retains two decimal places of precision. Then, each ratio value is merged with the corresponding node budget cost. The Zip structure is used to encapsulate the ratio value array and the budget value array into a composite key-value pair. The data frame merge function is called to align the combined data structure according to the node identifier, and finally the node resource density set is output.

[0065] Based on the node resource density set, a hierarchical dependency parsing method is used to extract the construction sequence and resource conflict relationship between nodes. First, the task flow table is called to read the front and back construction relationship of each node. The dependency order is recorded in an ordered list and assigned a sequence number according to the index code. Then, the resource matrix cross-comparison method is used to perform resource conflict identification. A Boolean matrix is constructed for all node resource requirements according to the resource type. The matrix dimension is the number of nodes × the number of resource types. Each item in the matrix indicates whether the node occupies a resource type. A column-by-column cross-logical AND operation is performed on all node combinations. If the same resource identification position is true, it is recorded as a resource conflict and a conflict identification set is generated. Adjacent edges are constructed for the two types of dependency relationships according to the sequential dependency number and conflict identification. The graph structure constructor is used to add sequential edges and conflict edges. The nodes are set as vertices and the dependencies are set as edge attributes. Finally, a construction task association relationship graph is output.

[0066] The specific steps to generate the comprehensive evaluation matrix of construction critical path are as follows:

[0067] Based on the construction task association graph, a graph convolutional network is used to extract node budget costs, resource intensity, construction sequence relationships, and resource conflict relationships. Standard normalization processing is performed on node budget costs and resource intensity to form a unified node indicator value set and construct a node comprehensive indicator matrix.

[0068] Based on the node comprehensive index matrix, the node order relationship and conflict relationship are extracted, the node sorting number is used to match the order dependency, the resource conflict ratio is used to determine the priority of the conflicting edge, the corresponding weight record table of the node and edge is generated, and the node edge weight matrix set is established;

[0069] Based on the node-edge weight matrix set, the node stage attributes, budget usage progress, and resource consumption level are extracted. The node comprehensive weight and stage budget completion rate are used to perform weighted accumulation. The accumulated results are used to perform descending sorting of node priorities to generate a comprehensive evaluation matrix for the construction critical path.

[0070] Based on the construction task association graph, the graph convolutional network method is used to extract the node budget cost, resource intensity, construction sequence relationship and resource conflict relationship. The node budget cost and resource intensity are input as node feature vectors. The input features are first normalized using the Min-Max normalization algorithm. The maximum and minimum values of each feature are extracted respectively. The tensor operation function is called to perform batch normalization on all node features according to the dimension. The normalized output range is set to 0 to 1. After normalization, the node budget cost and resource intensity are spliced by column to form a two-dimensional matrix structure. Each row represents a node and each column represents a normalized indicator value. The feature matrix is then input into the graph convolutional network using PyTorch. The Geometric library defines the model structure, setting the number of graph convolution layers to 2, the number of output channels per layer to 64, and the activation function to ReLU. The adjacency matrix is constructed using the directed edge structure of the task relationship graph. The GCNConv function is called to complete the adjacency feature propagation operation. Finally, a unified set of node indicator values is output. The node order and resource conflict relationship are combined to construct a two-dimensional tensor to generate a node comprehensive indicator matrix.

[0071] Based on the node comprehensive indicator matrix, the weight matching and ratio priority function method is used to extract the node sequence relationship and conflict relationship. First, the task sequence dependency edges are numbered and matched, and the node index numbers are extracted and arranged according to the topological sorting rules. The structured mapping function is used to convert the sequence relationship into the sequence edge weight. The sequence edge weight value is set to the inverse of the sorting number difference. The conflict weight matrix is constructed for the resource conflict relationship. The resource conflict ratio is used as the weight generation standard. The conflict ratio is calculated by the ratio of the number of overlapping resource types of two nodes to the total number of resource types. Then, the tensor splicing function is used to align the sequence edge weight matrix and the conflict edge weight matrix according to the node index. A sparse matrix structure containing all edge and weight information is constructed. Finally, the corresponding weight record table of nodes and edges is output and a node edge weight matrix set is generated.

[0072] Based on the node-edge weight matrix set, the weighted integral scoring method is used to extract the node stage attributes, budget usage progress and resource consumption level, and a node attribute tensor is constructed. Each node corresponds to a set of numerical vectors containing the stage code, stage budget completion rate and stage resource consumption quantity. A unified weight standard is used to perform a weighted accumulation operation on each indicator. The node comprehensive weight is set to 0.5, the stage budget completion rate is set to 0.3, and the resource consumption level is set to 0.2. The dot multiplication function is called to perform weighted addition on the above values, and the weighted cumulative value is combined with the node identifier to form a node scoring table. Subsequently, a descending sorting algorithm is used to perform descending sorting on the node scores. The sorting structure outputs the final matrix according to the task node number, and finally a comprehensive evaluation matrix of the construction critical path is generated.

[0073] Graph convolutional network, according to the formula:

[0074]

[0075] Among them: H (l+1) represents the comprehensive indicator matrix of the l+1th layer node, σ represents the nonlinear activation function, represents the degree matrix of the construction task association graph containing self-loops, represents the adjacency matrix of the construction task association graph containing self-loops, α represents the weight coefficient of the node budget cost attribute, represents the normalized characteristic matrix of the budget cost standard of the l-th layer node, β represents the weight coefficient of the node resource intensity attribute, represents the standard normalized feature matrix of the resource intensity of the l-th layer node, W (l) Represents the trainable weight matrix of layer l;

[0076] Execution process: First, build a construction task association diagram based on the construction tasks and generate an adjacency matrix and in Add the identity matrix to introduce self-loop connections, and then calculate the degree matrix in The diagonal elements are the number of connections of the corresponding nodes. Then the budget cost data of each node is normalized to form the budget cost feature matrix The resource intensity data of each node is normalized to form a resource intensity feature matrix Based on historical engineering data, regression analysis methods are applied to calculate the contribution percentages of budget cost and resource intensity to the total project cost and project progress respectively. The contribution percentages are set as basic weights, and weight coefficients α and β are obtained after normalization. The combined feature matrix is calculated based on the weight coefficients. Then use the formula Perform convolutional feature propagation on the combined feature matrix and then add it to the trainable weight matrix W (l) Perform matrix multiplication and finally perform nonlinear transformation through the activation function σ to output the comprehensive indicator matrix of the l+1th layer node, which provides a basis for subsequent dynamic measurement of engineering cost and cost control decision-making.

[0077] The specific steps to generate a dynamic abnormal node association subgraph are:

[0078] Based on the comprehensive evaluation matrix of the construction critical path, the actual progress value, actual expenditure value, and actual resource usage value of the tasks corresponding to each node are extracted. The progress offset is calculated by subtracting the actual progress from the node's planned progress, the expenditure offset is calculated by subtracting the actual expenditure from the node's budgeted expenditure, and the resource difference rate is calculated by subtracting the actual usage from the node's planned resources. This generates a set of node difference rate indicators.

[0079] Based on the node difference rate indicator set, the nodes with progress offset, expenditure offset, and resource difference rate exceeding the preset value are screened. The corresponding sequential dependency edge relationships are extracted using node index matching. The sequential edges are screened and sorted and merged to generate an abnormal node dependency edge set.

[0080] Based on the abnormal node dependency edge set, connect the abnormal nodes and the dependency edges, arrange the nodes in order and number them according to the edge connection order, establish a bidirectional index table of the node set and the edge set, draw the subgraph connection relationship, and establish a dynamic abnormal node association subgraph;

[0081] Based on the comprehensive evaluation matrix of the construction critical path, a structured field difference calculation method is used to extract the actual progress value, actual expenditure value and actual resource usage value corresponding to each node task. Field mapping processing is performed on the actual data set. The node number is used as the primary key to establish a field index table. The progress, budget expenditure and resource usage field values in the original plan data of the node are read. Vectorized subtraction operations are used to subtract the actual progress from the node plan progress, the actual expenditure from the node budget expenditure, and the actual usage from the node plan resources. Node-by-node difference calculation is performed. The calculation results are retained to two decimal places and uniformly converted into percentage format. The output results of the resource difference calculation are set to an upper limit of 100 and truncation processing is performed. The above three types of difference data are spliced into a three-dimensional tensor structure with the node number as the index dimension and the three types of difference indicators as the feature dimensions. Finally, a set of node difference rate indicators is output.

[0082] Based on the node difference rate indicator set, a multi-condition filtering function method is used to screen out nodes whose progress offset, expenditure offset, and resource difference rate exceed the preset value. The preset threshold is set to 10. A column-by-column scanning mechanism is used to compare the difference indicator value of each node in turn. A Boolean index matrix is constructed to mark the nodes that exceed the threshold. Subsequently, the node number is used as the primary key to perform sequential dependency edge query, extract the predecessor or successor task edge information corresponding to the matching node, and use the edge vector merging function to integrate duplicate dependency edges. The edge set is then deduplicated. The edge attributes include the starting node, the ending node, and the edge type identifier. The merged data structure is set to a three-column matrix, and the final output is the abnormal node dependency edge set.

[0083] Based on the abnormal node dependency edge set, the adjacency matrix construction and index mapping method are used to establish a bidirectional reference structure of the node set and the edge set. First, the start node and end node indexes in all edge records are read, a node set list is constructed and numbered and sorted, a node sequence vector is generated according to the numbering sequence, the edge set is numbered with edge connection sequence, and the adjacency matrix constructor is called to generate a sparse connection matrix, in which the edge value field is set to the edge sequence number, the connection weight is set to 1, and the node number and edge number are used as row and column indexes to construct a bidirectional index table respectively. Then, the graph drawing function is called with the index table and the connection matrix as input, the visual graph layout algorithm is executed on the node and edge sets, a structured connection graph is drawn, and a dynamic abnormal node association subgraph is output.

[0084] The specific steps for generating a joint optimization adjustment plan for budget and construction period are as follows:

[0085] Based on the dynamic abnormal node association subgraph, a non-dominated sorting genetic algorithm is used to extract the remaining budget value, remaining duration value, and remaining resource value corresponding to the abnormal node. The remaining budget is calculated by subtracting the accumulated expenditure value from the initial budget value of the node, the remaining duration is calculated by subtracting the number of completed duration days from the initial duration value of the node, and the remaining resource amount is calculated by subtracting the node resource plan amount from the used amount. This generates a set of remaining resource amounts for the abnormal node.

[0086] Based on the remaining resource set of abnormal nodes, nodes with a remaining budget ratio lower than the budget threshold and nodes with a remaining duration ratio lower than the duration threshold are screened. High-priority nodes are screened based on the remaining resource ratio of the nodes. Node priorities are rearranged according to the combined value of budget and duration to generate a set of nodes with optimized budget and duration.

[0087] Based on the budget duration optimization node set, extract the node remaining budget, remaining duration, and remaining resources. Use the node budget remaining and node duration remaining to perform intersection screening, extract node combinations to form resource replenishment path solutions, organize the combination optimization data table, and establish a budget duration joint optimization adjustment plan.

[0088] Based on the dynamic abnormal node association subgraph, the non-dominated sorting genetic algorithm is used to extract the remaining budget, remaining duration, and remaining resource values corresponding to the abnormal nodes. For each node, its initial budget, cumulative expenditure, initial duration, completed days, resource plan number, and used resource number are read. The field-by-field calculation function is called to perform subtraction operations on the budget field to obtain the remaining budget, subtraction of days on the duration field to obtain the remaining duration, and difference operation on the resource field to obtain the remaining resource amount. Three columns of numerical vector arrays are constructed as the initial population feature input. During the execution of the non-dominated sorting genetic algorithm, the population size is set to 200, the crossover probability is 0.9, and the mutation probability is 0.1. The selection strategy adopts the tournament selection method, the crossover operation type is simulated binary crossover, the mutation operation adopts the polynomial mutation method, and the evolutionary round is set to 100. The corresponding nodes and their remaining budgets, durations, and resources in the current Pareto optimal solution are retained, and the final output is the remaining resource amount set of the abnormal nodes.

[0089] Based on the set of remaining resources of abnormal nodes, an interval screening and joint sorting algorithm is used to screen out nodes whose remaining budget ratio is lower than the budget threshold and nodes whose remaining duration ratio is lower than the duration threshold. The budget threshold is set to 20 and the duration threshold is set to 25. The remaining budget ratio and the remaining duration ratio are calculated item by item according to the node fields. A Boolean mask array is constructed to screen out the node sets that do not meet the budget and duration thresholds. Then, a sorting operation is performed on the remaining resource ratio field in descending order. A set of high-priority nodes with a remaining resource ratio greater than 80 is extracted. The screened nodes are formed into a two-dimensional vector according to the budget ratio and duration ratio. The Lexicographic sorting algorithm is used to rearrange the node priorities. During the sorting process, the budget ratio weight is set to 0.6 and the duration ratio weight is set to 0.4. Finally, a set of nodes with optimized budget and duration is output.

[0090] Based on the budget duration optimization node set, the intersection mapping and supplementary path construction algorithm are used to extract the node remaining budget, remaining duration and remaining resources. First, three sets of vector lists are constructed to store the remaining budget value, remaining duration value and remaining resource value of each node respectively. The intersection filtering function is called to perform double filtering on the budget remaining value and duration remaining value fields, and a set of nodes that meet both the budget greater than 50 and the duration greater than 40 are extracted. The screening results are matched with the remaining resource array, and the K-nearest neighbor resource complementary retrieval method is used to perform node combination. The margin difference calculation is performed on the resource type dimension, and node pairs with complementary differences less than 10 are selected to establish a combined path. The path structure is set to a directed edge structure, and each edge record contains the starting node, the target node and the resource difference value. Finally, all combined paths are sorted and the path structured table is output according to the resource type classification to generate a budget duration joint optimization adjustment plan.

[0091] Non-dominated sorting genetic algorithm, according to the formula:

[0092]

[0093] Where: f i represents the comprehensive remaining index value of the i-th abnormal node, w1 represents the importance weight coefficient of the remaining budget ratio, w2 represents the importance weight coefficient of the remaining duration ratio, w3 represents the importance weight coefficient of the remaining resource ratio, w4 represents the importance weight coefficient of the progress completion correction item, and RB i Indicates the remaining budget value of the i-th abnormal node, IB i represents the total initial budget of the i-th abnormal node, RD i Indicates the remaining duration value of the i-th abnormal node, ID i represents the initial duration of the i-th abnormal node, RR i Indicates the remaining resource value of the i-th abnormal node, IR i represents the total planned resources of the ith abnormal node, PC i represents the progress completion percentage correction item of the i-th abnormal node;

[0094] Execution process: First extract the remaining budget value RB of the dynamic abnormal node i , the remaining duration value RD i , remaining resource value Synchronously extract the initial budget value IB corresponding to the node i , the initial value of the construction period and the initial value of the resource plan, the remaining budget value is obtained by subtracting the cumulative value of expenditure from the initial value of the node budget, the remaining construction period value is obtained by subtracting the number of completed construction days from the initial value of the node construction period, and the remaining resource value is obtained by subtracting the used amount from the node resource plan number, the remaining budget ratio, the remaining construction period ratio, and the remaining resource ratio are calculated, and the node progress completion percentage is further extracted. After standardization, the progress correction item is formed, and historical project data is collected. The entropy weight method is used to calculate the characteristic information entropy of the remaining budget ratio, the remaining construction period ratio, the remaining resource ratio and the progress correction item. According to the information entropy, the importance weight coefficients w1, w2, w3, and w4 of each indicator are reversely determined, and the weight coefficients are normalized so that their sum is 1. Subsequently, each remaining ratio indicator and progress correction item are multiplied by the corresponding weight coefficient, and finally the comprehensive remaining indicator value of the abnormal node is calculated to provide data support for subsequent dynamic cost estimation and cost control.

[0095] The specific steps for generating a dynamic project budget and schedule optimization execution list are as follows:

[0096] Based on the budget duration joint optimization and adjustment plan, extract the remaining budget value, remaining duration value, and resource change rate value of each node. Use the remaining budget value and the project approved budget value to perform the ratio calculation and record the budget proportion. Use the remaining duration value and the approved duration value to perform the ratio calculation and record the duration tolerance. Use the resource change rate value and the node standard resource ratio to perform the deviation rate calculation to generate the budget duration ratio indicator set.

[0097] Based on the budget duration ratio indicator set, the node set whose budget ratio is less than the set budget lower limit and whose duration tolerance is less than the set duration lower limit is screened. The node set is sorted in ascending order by the budget deviation rate and the budget priority nodes are screened. The node set is sorted in ascending order by the duration deviation rate and the duration priority nodes are screened. The intersection set of the budget priority nodes and the duration priority nodes is extracted to generate a budget duration correction node combination.

[0098] Based on the combination of budget duration correction nodes, the remaining budget value, remaining duration value, and resource change rate value of the intersection node are extracted. The budget adjustment amount is generated by adding up the remaining budget values of the nodes, and the duration adjustment amount is generated by adding up the remaining duration values of the nodes. The budget supplement and duration supplement are integrated according to the cost adjustment path corresponding to the resource change rate of each node to establish a dynamic budget and duration optimization execution list for the project;

[0099] Based on the budget-duration joint optimization and adjustment plan, the proportion conversion and difference evaluation algorithm is used to extract the remaining budget value, remaining duration value and resource change rate value of each node, read the original approved budget value and approved duration value of the node, call the vectorized ratio calculation function to divide the remaining budget value of each node by the corresponding approved budget value, retain the result in percentage format and record it as the budget ratio, use the same method to divide the remaining duration value by the corresponding approved duration value, record it as the duration tolerance, then perform the difference operation on the resource change rate value and the node standard resource ratio and divide it by the standard ratio, the result is defined as the deviation rate and retains three decimal places, the budget ratio, duration tolerance and resource deviation rate of all nodes are combined into a three-column vector table according to the index, and finally output the budget duration ratio indicator set;

[0100] Based on the budget duration ratio indicator set, the node set is screened and rearranged using threshold screening and ascending sorting algorithms. The budget lower limit is set to 30, and the duration lower limit is set to 35. The budget ratio field is compared node by node and the node set with a value less than the budget lower limit is screened. The duration tolerance field is compared node by node and the node set with a value less than the duration lower limit is screened. The budget screening node set is sorted in ascending order by the resource deviation rate field, and the top 40% of nodes are taken to form the budget priority node set. The duration screening node set is sorted in ascending order by the resource deviation rate field, and the top 40% of nodes are taken to form the duration priority node set. Subsequently, the set intersection function is used to obtain the intersection result of the two node sets, and the node number and the corresponding budget ratio, duration tolerance, and resource deviation rate indicators are retained. Finally, the budget duration correction node combination is output.

[0101] Based on the combination of budget duration correction nodes, the numerical aggregation and path matching fusion algorithm is used to extract the remaining budget value, remaining duration value and resource change rate value of the intersection node. The remaining budget values of all intersection nodes are added to generate the budget adjustment amount, and all remaining duration values are added to generate the duration adjustment amount. The resource change rate field participates in the cost path mapping, and the cost path matching function is called to combine the resource deviation rate with the unit price influencing factor. The budget adjustment node and duration extension node in the corresponding resource supplement path are found, and the budget increase path and duration increase path are constructed in sequence according to the node index. Each path records the resource type, node number and increase ratio. Finally, all path data are summarized and a standardized structured list format is generated to output the project dynamic budget and duration optimization execution list.

[0102] The remaining budget value and the project approved budget value are calculated in proportion and the budget ratio is recorded. The remaining budget value of the current node is calculated by ratio with the project approved budget value. The calculated ratio is used to reflect the current budget execution status of the node and to determine whether the budget execution is in line with the predetermined plan. The budget ratio is recorded proportionally.

[0103] See also Figure 2 and Figure 3 An intelligent engineering cost dynamic calculation and cost control system is provided. The intelligent engineering cost dynamic calculation and cost control system is used to implement the above-mentioned intelligent engineering cost dynamic calculation and cost control method. The system includes:

[0104] Construction task data acquisition module: Based on construction task information, it extracts the engineering quantity, unit price, construction period, and resource requirements of the foundation unit, main unit, and decoration unit, calculates the cost of each node, and generates the resource density of the node through proportional coefficient calculation. It also combines the construction sequence and resource conflict relationship to generate a construction task association diagram;

[0105] Association graph construction module: Based on the construction task association graph, a graph convolutional network is used to extract node cost, resource intensity, sequence relationship, and resource conflict information. The dependency relationships between nodes are calculated and dependency judgment data is generated. This is followed by an edge weight matrix. The weighted feature accumulation and sorting are performed through the progress synchronization table and the cost prediction unit to generate a comprehensive evaluation matrix for the construction critical path.

[0106] Construction critical path assessment module: Based on the comprehensive construction critical path assessment matrix, it extracts the actual progress, expenditure, and resource usage of tasks, calculates the progress difference, expenditure deviation, and resource excess rate of nodes, filters out abnormal nodes, extracts dependent edge sets, and generates a dynamic abnormal node association subgraph;

[0107] Budget and duration optimization module: Based on the dynamic abnormal node association subgraph, the module uses a non-dominated sorting genetic algorithm to extract the remaining budget, duration, and resource values of each node, calculate the execution status of the budget and duration, select the optimized nodes, and combine resource allocation with remaining resource accounting to generate a joint optimization adjustment plan for the budget and duration.

[0108] Budget and schedule execution list module: Based on the budget and schedule joint optimization adjustment plan, extract the remaining budget, schedule and resource change rate, calculate the budget proportion and schedule tolerance, screen out the correction combination, combine the cost adjustment path, and generate the project dynamic budget and schedule optimization execution list.

[0109] The above are merely preferred embodiments of the present invention and do not limit the present invention in any other form. Any technician familiar with the profession may use the technical content disclosed above to change or modify it into an equivalent embodiment with equivalent changes and apply it to other fields. However, any simple modification, equivalent change and modification made to the above embodiment based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall still fall within the scope of protection of the technical solution of the present invention.

Claims

1. An intelligent engineering cost dynamic calculation and cost control method, characterized in that: The following steps are involved: S1: Based on the construction task information, the project quantity, unit price, construction period and resource requirements of the foundation, main structure and decoration units are extracted, the node cost is calculated, the node resource density is generated through proportional coefficient calculation, the construction sequence and resource conflict relationship are combined, and a construction task association diagram is generated; S2: Based on the construction task association graph, a graph convolutional network is used to extract node costs, resource intensity, sequence relationships, and resource conflicts to generate a comprehensive node impact matrix. An edge weight matrix is generated through dependency judgment. Combined with the schedule synchronization table and the cost prediction unit, feature weighted accumulation and sorting are performed to obtain a comprehensive construction critical path evaluation matrix. S3: Based on the comprehensive evaluation matrix of the construction critical path, the actual progress, expenditure, and resource usage of the task are extracted, the node progress difference, expenditure deviation, and resource excess rate are calculated, abnormal nodes are screened, the dependent edge set is extracted, and a dynamic abnormal node association subgraph is generated; S4: Based on the dynamic abnormal node association subgraph, a non-dominated sorting genetic algorithm is used to extract the remaining budget, duration, and resource values, perform budget and duration judgment, screen optimization nodes, and generate a budget and duration joint optimization adjustment plan by combining resource allocation and remaining resource accounting; S5: Based on the budget-construction-period joint optimization and adjustment plan, extract the remaining budget, construction period and resource change rate, calculate the budget proportion and construction period tolerance, screen the correction combination, combine the cost adjustment path, and generate the project dynamic budget and construction period optimization execution list.

2. The intelligent engineering cost dynamic calculation and cost control method according to claim 1 is characterized in that: The specific steps for generating the construction task association diagram are: Based on the construction task information, the engineering quantity, unit price, construction period, and resource requirements of the foundation unit, main unit, and decoration unit are extracted. The engineering quantity and unit price values are multiplied one by one to form the node budget cost. The node budget cost and node identification index are used to organize the mapping relationship and establish the node budget cost set. Based on the node budget cost set, the engineering quantity and resource requirement quantity corresponding to each node are extracted. The node engineering quantity is divided by the resource requirement quantity to perform a ratio calculation. The ratio calculation value is then associated with the node budget cost to form a combined data set, generating a node resource intensity set. Based on the node resource density set, the construction sequence and resource conflict relationship between nodes are extracted, the dependency relationship is classified using sequence numbering, and the conflict relationship identification is formed by cross-comparison of resource occupancy quantity. The sequence dependency edge and the conflict dependency edge are connected to establish the construction task association relationship diagram.

3. The intelligent engineering cost dynamic calculation and cost control method according to claim 1 is characterized in that: The specific steps for generating the comprehensive evaluation matrix of the construction critical path are as follows: Based on the construction task association graph, a graph convolutional network is used to extract the node budget cost, resource intensity, construction sequence relationship, and resource conflict relationship. The node budget cost and resource intensity are normalized using standard processing to form a unified node indicator value set and construct a node comprehensive indicator matrix. Based on the node comprehensive indicator matrix, the node order relationship and conflict relationship are extracted, the node sorting number is used to match the order dependency, the resource conflict ratio is used to determine the priority of the conflicting edge, the corresponding weight record table of the node and edge is generated, and the node edge weight matrix set is established; Based on the node-edge weight matrix set, the node stage attributes, budget usage progress, and resource consumption level are extracted, and weighted accumulation is performed using the node comprehensive weight and the stage budget completion rate. The accumulated results are used to perform descending sorting of the node priorities to generate a comprehensive evaluation matrix for the construction critical path.

4. The intelligent engineering cost dynamic calculation and cost control method according to claim 3 is characterized in that: The graph convolutional network, according to the formula: Among them: H (l+1) represents the comprehensive indicator matrix of the l+1th layer node, σ represents the nonlinear activation function, represents the degree matrix of the construction task association graph containing self-loops, represents the adjacency matrix of the construction task association graph containing self-loops, α represents the weight coefficient of the node budget cost attribute, represents the normalized characteristic matrix of the budget cost standard of the l-th layer node, β represents the weight coefficient of the node resource intensity attribute, represents the standard normalized feature matrix of the resource intensity of the l-th layer node, W (l) Represents the trainable weight matrix of layer l.

5. The intelligent engineering cost dynamic calculation and cost control method according to claim 1 is characterized in that: The specific steps of generating the dynamic abnormal node association subgraph are as follows: Based on the comprehensive evaluation matrix of the construction critical path, the actual progress value, actual expenditure value, and actual resource usage value of the tasks corresponding to each node are extracted. The progress offset is calculated by subtracting the actual progress from the node's planned progress, the expenditure offset is calculated by subtracting the actual expenditure from the node's budgeted expenditure, and the resource difference rate is calculated by subtracting the actual usage from the node's planned resources, thereby generating a node difference rate indicator set. Based on the node difference rate indicator set, the nodes with progress offset, expenditure offset, and resource difference rate exceeding the preset value are screened, the corresponding sequential dependency edge relationships are extracted using node index matching, the sequential edges are screened and sorted and merged, and a set of abnormal node dependency edges is generated; Based on the abnormal node dependent edge set, the abnormal nodes and dependent edges are connected, the nodes are arranged in order and numbered according to the edge connection order, a bidirectional index table of the node set and the edge set is established, the subgraph connection relationship is drawn, and a dynamic abnormal node associated subgraph is established.

6. The intelligent engineering cost dynamic calculation and cost control method according to claim 1 is characterized in that: The specific steps for generating the budget duration joint optimization adjustment plan are as follows: Based on the dynamic abnormal node association subgraph, a non-dominated sorting genetic algorithm is used to extract the remaining budget value, remaining duration value, and remaining resource value corresponding to the abnormal node, calculate the remaining budget by subtracting the accumulated expenditure value from the initial budget value of the node, calculate the remaining duration by subtracting the number of completed duration days from the initial duration value of the node, and calculate the remaining resource amount by subtracting the node resource plan amount from the used amount, thereby generating a remaining resource amount set for the abnormal node; Based on the remaining resource amount set of the abnormal nodes, the nodes whose remaining budget ratio is lower than the budget threshold and the nodes whose remaining duration ratio is lower than the duration threshold are screened, the high-priority nodes are screened according to the node resource remaining ratio, the node priorities are rearranged according to the combined value of budget and duration, and a budget-time optimized node set is generated; Based on the budget duration optimization node set, the node remaining budget, remaining duration, and remaining resource amount are extracted, and intersection screening is performed using the node budget remaining amount and the node duration remaining amount. The node combination is extracted to form a resource supplement path plan, the combination optimization data table is sorted out, and a budget duration joint optimization adjustment plan is established.

7. The intelligent engineering cost dynamic calculation and cost control method according to claim 6 is characterized in that: The non-dominated sorting genetic algorithm is based on the formula: Where: f i represents the comprehensive remaining index value of the i-th abnormal node, w1 represents the importance weight coefficient of the remaining budget ratio, w2 represents the importance weight coefficient of the remaining duration ratio, w3 represents the importance weight coefficient of the remaining resource ratio, w4 represents the importance weight coefficient of the progress completion correction item, and RB i Indicates the remaining budget value of the i-th abnormal node, IB i represents the total initial budget of the i-th abnormal node, RD i Indicates the remaining duration value of the i-th abnormal node, ID i represents the initial duration of the i-th abnormal node, RR i Indicates the remaining resource value of the i-th abnormal node, IR i represents the total planned resources of the ith abnormal node, PC i Represents the progress completion percentage correction item of the i-th abnormal node.

8. The intelligent engineering cost dynamic calculation and cost control method according to claim 1 is characterized in that: The specific steps for generating the project dynamic budget and construction period optimization execution list are as follows: Based on the budget duration joint optimization and adjustment plan, the remaining budget value, remaining duration value, and resource change rate value of each node are extracted, the remaining budget value and the project approved budget value are used to perform a ratio calculation and record the budget proportion, the remaining duration value and the approved duration value are used to perform a ratio calculation and record the duration tolerance, and the resource change rate value and the node standard resource ratio are used to perform a deviation rate calculation to generate a budget duration ratio indicator set; Based on the budget-to-duration ratio indicator set, a node set whose budget ratio is less than a set budget lower limit and whose duration tolerance is less than a set duration lower limit is selected, the node set is sorted in ascending order by the budget deviation rate and budget priority nodes are selected, the node set is sorted in ascending order by the duration deviation rate and duration priority nodes are selected, an intersection set of the budget priority nodes and the duration priority nodes is extracted, and a budget-to-duration correction node combination is generated; Based on the budget duration correction node combination, the remaining budget value, remaining duration value, and resource change rate value of the intersection node are extracted, and the budget adjustment amount is generated by adding up the remaining budget values of the nodes, and the duration adjustment amount is generated by adding up the remaining duration values of the nodes. The budget supplement and duration supplement are integrated according to the cost adjustment path corresponding to the resource change rate of each node to establish a dynamic budget and duration optimization execution list for the project.

9. The intelligent engineering cost dynamic calculation and cost control method according to claim 8 is characterized in that: The remaining budget value and the project approved budget value are calculated in proportion and the budget ratio is recorded. The remaining budget value of the current node is calculated by ratio with the project approved budget value. The calculated ratio is used to reflect the current budget execution status of the node, to determine whether the budget execution is in line with the predetermined plan, and the budget ratio is recorded proportionally.

10. An intelligent engineering cost dynamic calculation and cost control system, characterized in that: According to any one of claims 1 to 9, the intelligent engineering cost dynamic calculation and cost control method comprises: Construction task data acquisition module: Based on construction task information, it extracts the engineering quantity, unit price, construction period, and resource requirements of the foundation unit, main unit, and decoration unit, calculates the cost of each node, and generates the resource density of the node through proportional coefficient calculation. It also combines the construction sequence and resource conflict relationship to generate a construction task association diagram; Association graph construction module: Based on the construction task association graph, a graph convolutional network is used to extract node cost, resource density, sequence relationship, and resource conflict information, calculate the dependency relationship between each node, and generate dependency judgment data. This module then generates an edge weight matrix. The progress synchronization table and cost prediction unit perform weighted feature accumulation and sorting to generate a comprehensive evaluation matrix for the construction critical path. Construction critical path assessment module: Based on the construction critical path comprehensive assessment matrix, it extracts the actual progress, expenditure, and resource usage of the task, calculates the progress difference, expenditure deviation, and resource excess rate of the node, filters out abnormal nodes, extracts the dependent edge set, and generates a dynamic abnormal node association subgraph; Budget and duration optimization module: Based on the dynamic abnormal node association subgraph, a non-dominated sorting genetic algorithm is used to extract the remaining budget, duration, and resource values of each node, calculate the execution status of the budget and duration, select the optimized nodes, and combine resource allocation with remaining resource accounting to generate a budget and duration joint optimization adjustment plan; Budget and schedule execution list module: Based on the budget and schedule joint optimization adjustment plan, extract the remaining budget, schedule and resource change rate, calculate the budget proportion and schedule tolerance, screen out the correction combination, combine the cost adjustment path, and generate the project dynamic budget and schedule optimization execution list.

Images