Project four-control digital monitoring optimization method and system based on dynamic model

Abstract

The present application relates to the technical field of project management, and particularly relates to a project four-control digital monitoring optimization method and system based on a dynamic model.The present application proposes the following scheme: by acquiring plan progress data, real-time execution data and resource allocation data of a project process, using a time series prediction model to predict and analyze the real-time execution data, a progress deviation event is generated.Based on the updated progress dynamic model, a four-dimensional occupancy voxel set is constructed and sparse storage operation is performed to realize space-time conflict detection between processes, calculate a space-time conflict coefficient, and generate a conflict event library.The present application can realize dynamic monitoring of engineering progress, intelligent detection of conflicts and optimal scheduling of resources, and improve the real-time performance, accuracy and intelligent level of progress control.

Description

Technical Field

[0001] This invention relates to the field of engineering project management technology, specifically to a digital monitoring and optimization method and system for four-control systems of a project based on a dynamic model. Background Technology

[0002] In modern engineering project management, the "four controls" (i.e., schedule, cost, safety, and quality) are key factors in ensuring project success. Traditional project management methods largely rely on static plans and manual reports, leading to lagging and inaccurate monitoring of project progress and a lack of real-time feedback. When faced with complex project environments, especially when processes are interdependent, resources are scarce, and construction sites change rapidly, this traditional approach is prone to schedule deviations, cost overruns, safety hazards, and quality problems. It fails to make timely and effective adjustments and optimizations, severely impacting the overall project management efficiency and execution effectiveness.

[0003] In recent years, with the development of information technology, many engineering projects have gradually introduced digital monitoring methods based on Building Information Modeling (BIM), the Internet of Things (IoT), and big data analytics. These technologies can collect data from the project site in real time, providing real-time progress and resource allocation information for certain processes. However, existing digital monitoring systems lack a unified dynamic model, making it difficult to achieve coordination between various control objectives in both time and space. For example, schedule delays may lead to redundant resource investment and increased costs, overlapping construction operations may cause safety hazards, and rework due to quality defects will further affect schedule and cost, making it difficult to achieve overall optimization in a multi-objective, multi-constraint environment.

[0004] Especially in complex construction environments, conflicts between work processes, resources, and time and space are becoming increasingly apparent. These conflicts often lead to project delays, cost overruns, and safety accidents. Existing monitoring systems typically rely on static judgments based on experience or simple rules when detecting conflicts, making it impossible to handle complex spatiotemporal conflicts and achieve comprehensive real-time optimization.

[0005] Therefore, there is an urgent need to propose a new technical solution that can unify and analyze data on schedule, cost, safety, and quality based on dynamic models, enabling real-time monitoring, intelligent prediction, and optimization decision-making for these aspects. By introducing dynamic models, the evolution of engineering projects in both spatiotemporal and resource consumption dimensions can be more comprehensively depicted, overcoming the limitations of existing methods and improving the digitalization and intelligence of project management. Summary of the Invention

[0006] The technical problem this invention addresses is the shortcomings of existing technologies. It provides a project four-control digital monitoring and optimization method and system based on a dynamic model. By acquiring planned progress data, real-time execution data, and resource allocation data for project processes, a time-series prediction model is used to predict and analyze the real-time execution data, comparing it with planned progress data to generate schedule deviation events. Based on the updated dynamic progress model, a four-dimensional occupancy voxel set is constructed and sparse storage operations are performed to detect spatiotemporal conflicts between processes, calculate the spatiotemporal conflict coefficient, and generate a conflict event library. A multi-objective optimization model combined with heuristic algorithms is used to dynamically schedule resource allocation, generating resource optimization schemes to achieve a comprehensive balance between schedule delays, resource idleness, and construction conflicts. This invention enables comprehensive digital monitoring of four dimensions of a project, intelligent conflict detection, and optimal resource scheduling, significantly improving the real-time performance, accuracy, and intelligence level of project management.

[0007] To achieve the above objectives, the present invention provides the following technical solution:

[0008] A project four-control digital monitoring optimization method based on dynamic models is applied to the schedule, cost, safety, and quality management of engineering projects. The method includes:

[0009] Acquire planned progress data, real-time execution data, and resource allocation data for each process in an engineering project;

[0010] Perform time-series predictive analysis on the real-time execution data, compare it with the planned progress data, calculate the process progress deviation, and generate a progress deviation event.

[0011] The schedule deviation event is input into the schedule dynamic model to update the schedule dynamic model, wherein the schedule dynamic model is used to represent the dynamic state of the process in time and space.

[0012] Based on the updated progress dynamic model, inter-process conflict detection is performed to generate a conflict event library. The inter-process conflict detection is based on the updated progress dynamic model to detect the spatiotemporal overlap between processes and to quickly calculate the spatiotemporal conflict coefficient through a sparse storage structure.

[0013] Based on the conflict event database, resource scheduling is optimized using an optimization algorithm to generate a resource optimization scheme.

[0014] Perform time-series predictive analysis on the real-time execution data, compare it with the planned schedule data, calculate the process schedule deviation, and generate schedule deviation events, including:

[0015] The real-time execution data is preprocessed.

[0016] The preprocessed real-time execution data is input into the time series prediction model to generate progress prediction data for future periods;

[0017] The progress prediction data is compared with the planned progress data step by step to obtain the progress deviation data.

[0018] Based on the schedule deviation data, the dependencies between processes are identified by combining the process dependency graph, the deviation-related processes are determined, and a set of deviation-related processes is obtained, wherein the process dependency graph is obtained by the critical path algorithm;

[0019] Based on the schedule deviation data and the set of processes associated with the deviation, a schedule deviation event is generated.

[0020] The schedule forecast data is compared with the planned schedule data step by step to obtain schedule deviation data, including:

[0021] The predicted completion time of the progress forecast data and the planned completion time of the planned progress data are obtained.

[0022] The difference between the predicted completion time and the planned completion time of the process is calculated to obtain the progress deviation between the predicted progress data and the planned progress data.

[0023] The schedule deviation is compared with a preset deviation threshold. When the schedule deviation exceeds the deviation threshold, the corresponding process is marked as having a schedule deviation, and the process with the schedule deviation is obtained.

[0024] Based on the schedule deviation process and the corresponding schedule deviation amount, schedule deviation data is generated.

[0025] The schedule deviation event is input into the schedule dynamic model, and the schedule dynamic model is updated, including:

[0026] The schedule deviation process and the deviation-related process in the schedule deviation event are mapped to the process nodes in the schedule dynamic model, wherein the process nodes include time attributes and space attributes;

[0027] Update the time attributes of the process node based on the schedule deviation amount in the schedule deviation event, wherein the time attributes include start time, duration and end time;

[0028] The progress dynamic model is updated based on the adjusted time and space attributes of the process nodes.

[0029] The process of detecting inter-process conflicts and generating a conflict event library based on the updated progress dynamic model includes:

[0030] The process nodes of the progress dynamic model are mapped to a four-dimensional coordinate system containing spatial and temporal attributes, generating a four-dimensional voxel set of the process.

[0031] The four-dimensional occupied voxel set is written into a sparse storage structure, and a coarse screening is performed based on the four-dimensional boundary between processes to determine potentially overlapping process pairs.

[0032] Based on the overlapping process pairs, voxel set intersection operation is performed through sparse storage structure to obtain spatiotemporally overlapping voxels between processes, and a spatiotemporally overlapping voxel set is generated.

[0033] Based on the spatiotemporal overlap voxel set, the spatial overlap and temporal overlap between processes are calculated, and the spatiotemporal conflict coefficient between processes is generated.

[0034] The spatiotemporal conflict coefficient is compared with a preset conflict threshold to generate a conflict event library.

[0035] Based on the aforementioned spatiotemporal overlap voxel set, the spatial overlap and temporal overlap between processes are calculated, and the spatiotemporal conflict coefficient between processes is generated, including:

[0036] The spatial coordinates of the process are extracted from the four-dimensional occupied voxel set to obtain the spatial occupied voxel set of the process, and the spatial coordinates of the process overlap are extracted from the spatiotemporal overlapping voxel set to obtain the overlapping spatial voxel set between processes.

[0037] The spatial volume between two processes is compared to obtain the minimum value of the spatial volume between the two processes. The overlapping spatial volume is divided by the minimum value to obtain the spatial overlap. The spatial volume is the number of elements in the voxel set occupied by each process, and the overlapping spatial volume is the number of elements in the voxel set of the overlapping space between processes.

[0038] Determine the minimum and maximum time values ​​of each process from the four-dimensional voxel set of the process, and calculate the process duration;

[0039] Determine the minimum and maximum time values ​​from the spatiotemporal overlap voxel set, calculate the overlap duration, compare the process durations between two processes to obtain the shortest process duration between the two processes, and divide the overlap duration by the shortest value to obtain the time overlap degree.

[0040] Weights are set based on spatial overlap and temporal overlap, and the spatiotemporal conflict coefficient between processes is generated by weighted summation.

[0041] The spatiotemporal conflict coefficient is compared with a preset conflict threshold to generate a conflict event database, including:

[0042] When the spatiotemporal conflict coefficient is greater than or equal to the preset conflict threshold, the corresponding process pair is marked as a conflict event;

[0043] When the spatiotemporal conflict coefficient is less than the preset conflict threshold, the process pair is marked as conflict-free;

[0044] A conflict event library is generated based on all conflict events.

[0045] Based on the conflict event database, resource scheduling is optimized using an optimization algorithm on the resource configuration data to generate a resource optimization scheme, including:

[0046] Based on the conflict event database, obtain the corresponding work process pairs and corresponding resource configuration data, and optimize resource scheduling;

[0047] A multi-objective optimization model is established with the objective functions of minimizing project delays and minimizing resource scheduling costs. The constraints of the multi-objective optimization model include process dependencies, resource supply limits, and construction safety distances.

[0048] Resource optimization schemes are generated by adjusting resource configuration data using heuristic optimization algorithms.

[0049] The process involves adjusting resource allocation data using a heuristic optimization algorithm to generate a resource optimization scheme. The heuristic optimization algorithm is a particle swarm optimization algorithm, which includes:

[0050] Initialize the particle swarm position and velocity, where the particle swarm position represents the resource optimization scheme and the velocity represents the adjustment range;

[0051] The fitness of each particle is calculated based on the fitness function, which is a weighted average of project delay, resource idle rate, and number of conflict events.

[0052] Update the individual particle's optimal position and the global optimal position;

[0053] The particle swarm is iteratively updated based on the velocity update formula and the position update formula.

[0054] When the number of iterations or fitness converges, the globally optimal resource optimization scheme is output.

[0055] A project four-control digital monitoring and optimization system based on a dynamic model, the system comprising:

[0056] The data acquisition module is used to acquire planned progress data, real-time execution data, and resource configuration data;

[0057] The predictive analysis module is used to perform time-series predictive analysis on the real-time execution data, compare it with the planned progress data, calculate the process progress deviation, and generate a progress deviation event.

[0058] The dynamic model update module is used to maintain and update the dynamic schedule model, which includes process time attributes and spatial attributes.

[0059] The conflict detection module is used to input the schedule deviation events into the schedule dynamic model, update the schedule dynamic model, perform inter-process conflict detection, and generate a conflict event library.

[0060] The resource scheduling optimization module is used to optimize resource scheduling based on the conflict event database and through optimization algorithms to generate a resource optimization scheme.

[0061] The cost management module is used to realize the visual tracking and optimization control of costs throughout the entire construction life cycle, and to generate cost management reports;

[0062] The security management module is used to dynamically integrate the progress dynamic model with video surveillance to achieve video visualization management, and also integrate security reports with third-party systems.

[0063] The quality monitoring module is used to perform visualized management of project quality throughout the entire project lifecycle, based on schedule deviation events, conflict event databases, resource optimization schemes, cost management reports, and safety reports.

[0064] Compared with the prior art, the beneficial effects of the present invention are:

[0065] This invention introduces a time-series prediction model, enabling advance prediction of process progress based on real-time execution data. This allows for timely detection of potential deviations, preventing overall project delays due to accumulated delays. Deviation events are input into a dynamic model, which updates the temporal and spatial attributes of each process in real time, ensuring the model reflects the latest state of the construction site. By combining four-dimensional voxels with sparse storage, efficient spatiotemporal conflict detection between processes is achieved, improving detection speed while maintaining the accuracy of conflict identification, effectively preventing construction interference and safety hazards. Attached Figure Description

[0066] Other features, objects, and advantages of the invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings:

[0067] Figure 1 This is a schematic diagram of the project four-control digital monitoring optimization method based on dynamic model according to Embodiment 1 of the present invention;

[0068] Figure 2 This is a schematic diagram of a process progress deviation event in Embodiment 1 of the present invention;

[0069] Figure 3 This is a schematic diagram of the sparse storage structure in Embodiment 1 of the present invention;

[0070] Figure 4This is a schematic diagram of the spatiotemporal overlap voxel at time t in Embodiment 1 of the present invention;

[0071] Figure 5 This is a structural diagram of the project four-control digital monitoring optimization system based on a dynamic model, as shown in Embodiment 2 of the present invention. Detailed Implementation

[0072] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0073] Example 1

[0074] Please see Figure 1 The present invention provides an embodiment of a project four-control digital monitoring optimization method based on a dynamic model, the specific steps of which are as follows:

[0075] S1: Obtain the planned schedule data, real-time execution data, and resource allocation data of the work processes in the engineering project;

[0076] Specifically, the first step in project progress monitoring is the comprehensive collection and integration of data. Without complete basic data, subsequent predictions, deviation identification, and conflict detection will lack a reliable basis. This embodiment uses a multi-source data acquisition mechanism to obtain planned progress data, real-time execution data, and resource allocation data to construct a dynamic data input system covering the entire lifecycle.

[0077] In this embodiment, the planned progress data primarily originates from the project planning management system, and includes the planned start time, planned duration, planned completion time, and critical path information for each work process. Real-time execution data is automatically collected through sensors, construction logs, and the on-site management platform, including the actual start time of each work process, the real-time progress percentage, the current cumulative working hours, and equipment operating status. Resource allocation data includes the input and allocation of manpower, machinery, and materials at the construction site. For example, it includes the occupancy of key machinery for different work processes, the work shift arrangements, and the supply rhythm of major materials.

[0078] S2: Perform time-series predictive analysis on the real-time execution data, compare it with the planned progress data, calculate the process progress deviation, and generate a progress deviation event;

[0079] Specifically, during construction, the progress of each process is affected by various factors, such as equipment operating status, worker engagement, and material arrival. Relying solely on static plans for comparison often fails to reflect real-time on-site dynamics, leading to delays in progress deviation detection. Therefore, this embodiment employs a dynamic analysis method based on a time-series prediction model, comparing real-time execution data with planned progress data step-by-step to form a precise progress deviation identification mechanism.

[0080] In this embodiment, time-series modeling is first performed on the real-time execution data. By introducing prediction models such as recurrent neural networks, long short-term memory networks (LSTM), or temporal convolutional networks, and combining historical process data with the current state, process progress prediction results for future time periods are generated. The prediction results can reflect the completion trend of the process in the future time interval, avoiding random errors caused by single-point data.

[0081] Furthermore, the predicted completion time of each process is compared with the corresponding completion time in the planned schedule data, and the schedule deviation is calculated by difference. To ensure the robustness of the calculation results, this embodiment preferably adopts a multi-dimensional comparison strategy, that is, it simultaneously considers the differences in process start time, duration, and completion time, and forms a unified deviation index through a weighted mechanism. If the schedule deviation of a certain process exceeds a preset threshold, the process is automatically marked as having a schedule anomaly.

[0082] Preferably, this embodiment not only identifies deviations in a single process but also propagates these deviations by considering the dependencies between processes. For example, when a process on the critical path experiences a delay, the schedule deviation will be propagated to subsequent processes through the critical path algorithm, thereby generating a set of related deviation events.

[0083] S3: Input the schedule deviation event into the schedule dynamic model and update the schedule dynamic model;

[0084] Specifically, in construction projects, the actual execution of work processes often deviates from the plan and is identified as schedule deviation events. If deviation information cannot be fed back to the schedule dynamic model in a timely manner, the model will lose its dynamism and fail to accurately reflect the current state of the project. This embodiment uses an event-driven dynamic update mechanism, taking the detected schedule deviation events as input, to update the schedule dynamic model in real time.

[0085] In this embodiment, each schedule deviation event includes the deviation process, the deviation amount, and a set of associated processes. First, the deviation process and its related attributes are mapped to process nodes in the dynamic schedule model, and the time attributes of the process nodes (such as start time, duration, and end time) are corrected. Simultaneously, affected processes in the dependency path are adjusted based on the critical path algorithm to ensure that the dynamic schedule model reflects the logical connection and transmission effect between processes. For example, if a process in the critical path is delayed, the time parameters of subsequent processes will be adjusted synchronously to ensure the consistency of the overall schedule output by the model.

[0086] S4: Based on the updated progress dynamic model, perform inter-process conflict detection and generate a conflict event library;

[0087] Specifically, during engineering construction, complex spatiotemporal coupling relationships often exist between different work processes. For example, support operations, formwork installation, and concrete pouring within the same spatial area may overlap in time. If these processes are not detected and coordinated in a timely manner, they can easily lead to physical conflicts, resource contention, or even safety accidents. Traditional conflict detection methods typically rely on overall geometric comparison, which has high computational complexity and insufficient real-time performance. Therefore, this embodiment conducts inter-process conflict detection based on an updated progress dynamic model. By introducing four-dimensional voxel representation and a sparse storage mechanism, it achieves efficient conflict identification in both temporal and spatial dimensions.

[0088] In the specific implementation process, the temporal and spatial attributes of each process node in the updated progress dynamic model are first mapped to a four-dimensional coordinate system to obtain a four-dimensional occupied voxel set for the process. The four-dimensional occupied voxel set describes the distribution characteristics of the process in both three-dimensional spatial location and time dimension in a discrete unit manner. Subsequently, the four-dimensional occupied voxel set is written into a sparse storage structure, and preliminary screening is performed through inter-process boundary conditions to quickly identify potentially overlapping process pairs, avoiding the high overhead of global traversal.

[0089] Furthermore, for the selected overlapping process pairs, an intersection operation is performed under a sparse storage index to extract voxel units that overlap in both spatial and temporal dimensions, generating a spatiotemporally overlapping voxel set. The spatiotemporally overlapping voxel set can accurately reflect the intersection areas that may occur between processes in actual construction scenarios. For example, when rebar tying and formwork erection are arranged simultaneously in a certain area within the same time period, an overlap marker will be generated in the corresponding voxel set.

[0090] Furthermore, spatial overlap and temporal overlap are calculated separately: spatial overlap measures the degree of physical intersection of the work processes, while temporal overlap measures the degree of synchronization between the processes in terms of scheduling. By weighted fusion of the two, a spatiotemporal conflict coefficient between the processes is generated.

[0091] S5: Based on the conflict event database, optimize resource scheduling by using an optimization algorithm to generate a resource optimization scheme.

[0092] Specifically, in the actual construction process of engineering projects, conflicts between work processes often lead to resource contention, project delays, and even safety hazards. Relying solely on manual experience for scheduling can easily result in inconsistent optimization objectives, low resource utilization efficiency, and overall progress obstruction. Therefore, this embodiment optimizes resource scheduling based on a conflict event database, combining a multi-objective optimization model with heuristic algorithms to achieve dynamic coordination and global optimization of resource allocation.

[0093] In this embodiment, firstly, conflicting work process pairs and their corresponding resource requirements are extracted from a conflict event database. A multi-objective optimization model is then established, aiming to minimize total project delays, reduce resource idle rates, and avoid spatiotemporal conflicts. This multi-objective optimization model not only considers the time dependencies between work processes and construction safety distance constraints, but also incorporates realistic conditions such as resource supply limits to ensure the feasibility of the optimization results in actual construction.

[0094] Furthermore, to solve the aforementioned multi-objective optimization problem, this embodiment introduces a heuristic optimization algorithm (such as particle swarm optimization, genetic algorithm, or simulated annealing). Taking particle swarm optimization as an example, each particle represents a resource optimization scheme, the position parameter corresponds to the allocation method of resources among different processes, and the velocity parameter reflects the magnitude of resource adjustment. The particle swarm is iteratively optimized through a fitness function, which is composed of a weighted average of project delay, resource idle rate, and number of conflict events, thereby achieving a comprehensive balance between time, cost, and safety.

[0095] Please see Figure 2 The present invention provides a schematic diagram of process progress deviation events in Embodiment 1, wherein solid circles represent critical processes, dashed circles represent non-critical processes, solid arrows represent critical process dependencies, dashed arrows represent non-critical process dependencies, black circles represent deviation processes, and gray circles represent deviation-related processes.

[0096] The specific steps of S2 are as follows:

[0097] S2.1: Perform preprocessing operations on the real-time execution data;

[0098] Specifically, in construction site progress monitoring, real-time execution data comes from diverse sources, including construction status data collected by sensors, manually entered progress logs, and even auxiliary information from BIM models, video surveillance, or IoT devices. Due to the complexity of data sources and inconsistent collection frequencies, directly inputting raw data into the progress deviation analysis module can easily lead to abnormal noise, redundant information, or missing values, affecting the overall calculation accuracy. Therefore, before deviation detection, a systematic preprocessing operation is performed on the real-time execution data to ensure its integrity, accuracy, and consistency.

[0099] S2.2: Input the preprocessed real-time execution data into the time series prediction model to generate progress prediction data for future periods;

[0100] Specifically, in the dynamic monitoring of construction projects, relying solely on historical static plans is insufficient to effectively address the dynamics and uncertainties of the construction site. For example, equipment failures, changes in environmental conditions, or fluctuations in human resources can all lead to schedule deviations. Therefore, this embodiment, after preprocessing the real-time execution data, inputs it into a time-series forecasting model to generate schedule forecast data for future periods, thereby achieving early warning of potential schedule deviations.

[0101] In specific implementations, the time-series forecasting model can employ recurrent neural networks (RNNs), long short-term memory networks (LSTMs), gated recurrent units (GRUs), or attention-based time-series forecasting frameworks. The time-series forecasting model can capture the time dependence and nonlinear evolution characteristics of progress data, making it suitable for handling progress dynamics in complex construction environments. By inputting multi-dimensional real-time execution data (such as actual operation time, resource usage, and work completion rate), the model can generate predicted values ​​for the start time, duration, and end time of operations in future periods.

[0102] Furthermore, this embodiment introduces a multi-source data fusion mechanism in the prediction process. That is, in addition to real-time execution data, external constraint information from the construction site (such as weather forecasts, equipment status monitoring, and material supply conditions) can be used as auxiliary inputs to improve prediction accuracy. For example, when a high probability of rainfall is predicted for the next few days, the model will automatically adjust the predicted duration of earthwork operations to better reflect the actual situation.

[0103] S2.3: Compare the predicted progress data with the planned progress data step by step to obtain the progress deviation data;

[0104] Specifically, the key to construction progress monitoring lies in quantifying and identifying the deviations between the predicted results and the original plan, so as to take timely scheduling measures. Relying solely on a comparison of the overall construction period cannot reveal subtle deviations in specific processes, potentially leading to the accumulation of problems without being detected, thus affecting overall construction efficiency. Therefore, this embodiment performs a process-by-process comparative analysis of the progress forecast data and the planned progress data to achieve refined monitoring of progress deviations.

[0105] In the specific implementation, the first step is to map the process completion time, duration, and other attributes in the progress forecast data one-to-one with the corresponding time nodes in the planned progress data. This time node alignment ensures that the comparative analysis covers not only core processes on the critical path but also auxiliary processes on the non-critical path. Subsequently, using a differential calculation method, the difference between the forecast result and the planned time is calculated for each process, forming the process deviation data.

[0106] Furthermore, to avoid inaccurate deviation identification due to simple time differences, this embodiment introduces a dynamic threshold and a multi-dimensional comparison mechanism. For example, when calculating process deviations, not only the end time of the process is compared, but also the start time and duration; if a process is predicted to start earlier but its end is delayed due to an extended duration, the nature of the deviation can still be accurately identified. Simultaneously, the dynamic threshold is adjusted according to the importance of the process, the construction section it belongs to, and the critical path attributes: a stricter deviation threshold is set for critical processes, while ordinary processes are allowed some fluctuation to improve the rationality of the detection.

[0107] The specific steps of S2.3 are as follows:

[0108] S2.3.1: Obtain the predicted process completion time from the progress forecast data and the planned process completion time from the planned progress data;

[0109] Specifically, in dynamic monitoring of construction progress, the completion time of each process is a key indicator for measuring the consistency between actual progress and the plan. To accurately identify schedule deviations, the predicted completion time of each process is first extracted from the schedule forecast data and then matched with the planned completion time in the corresponding planned schedule data.

[0110] In practice, the predicted completion time of each work process is dynamically generated by a time-series forecasting model. This model, based on real-time execution data and multi-source external environmental data, reflects the progress trend of the construction site in future periods. The planned completion time, on the other hand, originates from the construction organization design or schedule baseline document, typically set before project commencement and used as a benchmark for assessing construction progress. The difference between the predicted and planned completion times constitutes the source of work process schedule deviations.

[0111] S2.3.2: Calculate the difference between the predicted completion time and the planned completion time of the process to obtain the progress deviation between the predicted progress data and the planned progress data;

[0112] Specifically, the time difference is first calculated by subtracting the predicted completion time from the planned completion time of each process. The difference result can be positive or negative; a positive value indicates that the process is lagging behind, a negative value indicates that the process is ahead of schedule, and a zero value indicates that the process is progressing as planned. To avoid misjudgment due to a single indicator, this embodiment further expands the dimensions of the difference calculation, that is, in addition to the completion time, it also calculates the difference in start time and the difference in duration, and integrates the three into a unified schedule deviation, which can more comprehensively reveal the nature of the process deviation: for example, if a process starts ahead of schedule but is still delayed in completion due to an extended duration, the deviation logic can be accurately captured.

[0113] S2.3.3: Compare the progress deviation with a preset deviation threshold. When the progress deviation exceeds the deviation threshold, mark the corresponding process as having a progress deviation and obtain the progress deviation process.

[0114] Specifically, during the dynamic monitoring of construction progress, while the progress deviation of each process can reflect the difference between the forecast and the plan, without proper filtering, some minor fluctuations may be misjudged as deviations, causing the model to be overly sensitive or triggering false alarms. Therefore, this embodiment introduces a threshold determination mechanism after obtaining the progress deviation through differential calculation. That is, by comparing the deviation with a preset deviation threshold process by process, only when the deviation exceeds the threshold is the corresponding process marked as having a progress deviation, thereby identifying the truly deviating process.

[0115] In practice, the deviation threshold can be flexibly set according to project characteristics and management requirements. For example, for critical path processes, the threshold is set lower to promptly identify and control potential delays; while for non-critical processes, the threshold can be appropriately relaxed to avoid unnecessary resource waste due to minor deviations in non-critical processes.

[0116] S2.3.4: Generate schedule deviation data based on the schedule deviation process and the corresponding schedule deviation amount.

[0117] Specifically, during the construction process, a single schedule deviation can only reflect local schedule anomalies. Without quantitative description and data-driven representation, it is difficult to support subsequent dynamic modeling and conflict detection. Therefore, this embodiment, after identifying the schedule deviation process, combines the corresponding schedule deviation amount to perform structured processing on the deviation information to generate schedule deviation data, thereby achieving accurate recording and traceable analysis of schedule anomalies.

[0118] S2.4: Based on the schedule deviation data, the process dependency graph is combined to identify the dependencies between processes, determine the deviation-related processes, and obtain a set of deviation-related processes, wherein the process dependency graph is calculated by the critical path algorithm between processes;

[0119] Specifically, deviations in construction progress are not only reflected in delays or advancements of individual processes, but can also propagate downstream through dependencies between processes, thereby affecting the overall construction rhythm. Therefore, this embodiment, after generating progress deviation data, further introduces a process dependency graph and uses a critical path algorithm to identify the logical relationships and path dependencies between processes, thereby determining the related processes that the deviation may affect.

[0120] In this embodiment, the critical path algorithm is first used to model the construction schedule, generating a process dependency graph. The process dependency graph uses process nodes as basic units, with edges representing the sequential constraints between processes, and labels the float times of critical path processes and non-critical path processes. Then, when a process is marked as a schedule deviation process, starting from the process node, the process dependency graph is recursively applied downstream to identify all process nodes potentially affected by delays or advancements, thus forming a set of deviation-related processes.

[0121] S2.5: Generate a schedule deviation event based on the schedule deviation data and the set of deviation-related processes.

[0122] Specifically, in dynamic monitoring of construction progress, simple progress deviation data can only reflect quantitative differences at the process level, and is insufficient to support subsequent model updates and conflict detection. Therefore, this embodiment generates structured progress deviation events based on deviation data and combined with a set of deviation-related processes, thereby realizing the transformation from numerical deviation to event-driven processes.

[0123] In the specific implementation process, the schedule deviation process is first bound to the corresponding schedule deviation amount to form the basic event unit of the deviation process. Then, based on the process dependency graph, it is extended downstream to identify the deviation-related processes affected by the schedule deviation process, and the deviation-related processes are included in the deviation events.

[0124] The specific steps for S3 are as follows:

[0125] S3.1: Map the schedule deviation process and the deviation-related process in the schedule deviation event to the process node in the schedule dynamic model, wherein the process node includes time attributes and space attributes;

[0126] Specifically, in construction progress monitoring, schedule deviations are often not isolated events, but rather caused by delays or advancements in a particular process, which then propagate to other related processes through logical dependencies between processes. Therefore, this embodiment, upon detecting a schedule deviation event, maps the deviation process and its associated processes to process nodes in the dynamic progress model, visually depicting the propagation path of the deviation within a unified dynamic model framework.

[0127] By simultaneously introducing temporal and spatial attributes, process nodes can not only reflect the progress status of the process itself, but also provide basic data for spatiotemporal conflict detection and resource scheduling optimization. For example, when the end time of a deviation process is delayed, it can be immediately identified with other adjacent processes based on spatial coordinates, and it can be determined whether there will be a risk of space occupation or temporal conflict.

[0128] S3.2: Update the time attribute of the process node according to the schedule deviation amount in the schedule deviation event, wherein the time attribute includes start time, duration and end time;

[0129] Specifically, in the dynamic modeling of construction progress, the time attribute of each work process node is a core parameter for measuring the consistency between the construction plan and the actual execution. When a schedule deviation event is detected, the amount of the schedule deviation is obtained and mapped to the time attribute of the corresponding work process node to dynamically update the model state, ensuring that the dynamic progress model reflects the latest situation on site in real time.

[0130] S3.3: Update the progress dynamic model based on the adjusted time and space attributes of the process nodes.

[0131] Specifically, after correcting the time attributes of the process nodes and related processes with schedule deviations, the schedule dynamic model needs to be updated as a whole to ensure that it can comprehensively and accurately reflect the latest status of the construction process. By simultaneously utilizing the time attributes (such as start time, duration, and end time) and spatial attributes (such as construction location and work scope) of the process nodes, the schedule dynamic model is dynamically iterated and updated, thereby achieving a spatiotemporal integrated expression of the construction site progress.

[0132] In the specific implementation process, the adjusted time attributes are first written into the dynamic progress model, updating the coordinate positions of the process nodes in the time dimension, and simultaneously correcting the dependencies between processes. For example, when the end time of a process is postponed, the earliest start time of the downstream process is automatically adjusted to ensure the logical consistency of the dependency chain. At the same time, spatial attributes are also mapped into the model in real time to recalculate the area occupied by the process in three-dimensional space, ensuring that the model can correctly reflect the operation position of the process in space.

[0133] Please see Figure 3A schematic diagram of the sparse storage structure in Embodiment 1 of the present invention, wherein the actual coordinates of the voxels are represented as (x... i y i , z i , t i ), x i y i , z i For three-dimensional spatial coordinates, t i For the time dimension, the address space is used to record the mapping relationship between the logical address of the voxel corresponding to the dependent process and the physical storage location. The numbers in the storage space represent the row number in the sparse matrix, reflecting the row position of the actual coordinates of the voxel in the storage space.

[0134] The specific steps for S4 are as follows:

[0135] S4.1: Map the process nodes to a four-dimensional coordinate system containing spatial and temporal attributes to generate a four-dimensional voxel set for the process;

[0136] Specifically, in the dynamic modeling process of engineering projects, a single-dimensional temporal or spatial representation is often insufficient to reveal the complex interactions between processes. Considering only the time dimension fails to capture the cross-interference between different processes on the construction site; relying solely on the spatial dimension makes it difficult to reflect the dynamic evolution of processes throughout the construction cycle. Therefore, this embodiment maps process nodes to a four-dimensional coordinate system containing three-dimensional spatial coordinates (x, y, z) and a time dimension (t) to form a four-dimensional voxel set for each process, achieving a unified expression of the dynamic attributes of the processes.

[0137] In practical implementation, the spatial scope (such as tunnel cross-sections and bridge pier areas) and start and end times of each process are first determined based on process design information and on-site monitoring data. Then, by discretizing the spatial dimension and setting a reasonable step size in the time dimension, the work area and time interval corresponding to each process are mapped to several four-dimensional voxel points. Each voxel unit contains spatial coordinates and a time identifier, thus characterizing the spatial and temporal distribution trajectory of the process at a microscopic level.

[0138] Furthermore, this embodiment introduces an adaptive voxel generation mechanism: in critical path processes or high-density construction areas, higher spatial resolution and finer time steps are used to ensure sensitivity and detection accuracy for potential conflicts; in non-critical processes or areas with a more relaxed work scope, a coarser voxel partitioning strategy can be used to balance storage overhead and computational complexity. Through differentiated modeling, both the accuracy of critical conflict detection and the overall computational efficiency can be improved.

[0139] S4.2: Write the four-dimensional occupied voxel set into a sparse storage structure, perform a coarse screening based on the four-dimensional boundary between processes, and determine potentially overlapping process pairs.

[0140] Specifically, during construction, procedures often overlap and intersect in both space and time, especially when resources are limited and procedures are numerous. To improve the efficiency and accuracy of conflict detection, this embodiment uses a four-dimensional occupancy voxel set for spatiotemporal overlap detection of procedures and optimizes the storage and computation process through a sparse storage structure.

[0141] First, the four-dimensional voxel set occupied by each process is written into a sparse storage structure. The four-dimensional voxel set contains the process's occupancy information in three-dimensional space (e.g., coordinates of the construction area) and time dimension. Since the number of processes is large and the space occupied by voxels is typically sparse, a sparse storage structure (such as a hash table or sparse matrix) is used to efficiently store the voxel information. The sparse storage structure only records the location and time information of non-empty voxel units, effectively reducing storage space and computational overhead.

[0142] Next, a preliminary screening is performed using the four-dimensional boundaries between processes. Specifically, the four-dimensional voxel set occupied by a process can be represented by boundaries in the spatial and temporal dimensions. The boundaries include the minimum bounding box of the process in space (such as the maximum range of length, width, and height) and the time interval (such as start time and end time). By comparing the four-dimensional boundaries of different processes, potentially overlapping process pairs can be quickly and initially screened out, avoiding the high computational burden caused by globally traversing all processes.

[0143] Furthermore, for the selected potentially overlapping process pairs, a precise spatiotemporal intersection operation is performed to calculate the actual spatiotemporal overlapping voxel set. This spatiotemporal overlapping voxel set will be used to generate spatiotemporal conflict coefficients and serve as input data for resource scheduling optimization.

[0144] Specifically, this embodiment effectively utilizes a sparse storage structure and coarse screening technology to reduce unnecessary computational burden, making the spatiotemporal conflict detection process more efficient and scalable. Simultaneously, it can accurately identify potential overlaps between processes, providing a precise basis for further conflict resolution and resource scheduling.

[0145] S4.3: Based on the overlapping process pairs, voxel set intersection operation is performed through sparse storage structure to obtain spatiotemporally overlapping voxels between processes, and a spatiotemporally overlapping voxel set is generated.

[0146] Specifically, in the actual processing, for the selected overlapping process pairs, the corresponding four-dimensional occupying voxel sets are extracted. Since the voxel data uses sparse storage, recording only the position and temporal attributes of non-empty voxel units, the four-dimensional occupying voxel sets of the overlapping process pairs can be quickly located using hash mapping or sparse indexing. Subsequently, element-wise matching and cross-referencing are performed on the four-dimensional occupying voxel sets of the two processes. If the coordinates of the same voxel coincide in both the three-dimensional spatial position and the temporal dimension, it is marked as a spatiotemporally overlapping voxel.

[0147] Furthermore, the sparse intersection calculation process maintains computational efficiency in large-scale construction scenarios, avoiding the exponential overhead of full-space traversal. It also ensures the accuracy of the detection results, guaranteeing that the final spatiotemporally overlapping voxel set accurately reflects the conflict zones between construction processes. For example, in tunnel construction, if process A (shotcrete operation) and process B (reinforcement installation) are both scheduled at the same cross-section and time segment, the overlapping voxels corresponding to the cross-section are marked in the sparse voxel intersection result, thus providing reliable input for spatiotemporal conflict coefficient calculation and resource optimization.

[0148] S4.4: Based on the spatiotemporal overlap voxel set, calculate the spatial overlap and temporal overlap between processes, and generate the spatiotemporal conflict coefficient between processes;

[0149] In this embodiment, based on the spatiotemporal overlapping voxel set, the spatial overlap and temporal overlap are calculated respectively, and a spatiotemporal conflict coefficient is generated by weighted fusion to achieve a comprehensive judgment in both spatial and temporal dimensions.

[0150] Specifically, during construction, different procedures often partially overlap spatially and temporally. For example, the formwork erection and rebar tying procedures in a certain area may cover the same cross-sectional area spatially, but their time overlaps only to a limited extent. If spatial overlap is used as the sole criterion, the risk of conflict may be overestimated; if temporal overlap is used as the sole criterion, the actual interference caused by competition for resources within the same area may be overlooked.

[0151] At the spatial level, three-dimensional coordinates are first extracted from the four-dimensional voxel set of each process to obtain the spatial voxel set of each process. Then, non-repeating three-dimensional coordinates are extracted from the spatiotemporally overlapping voxel set to obtain the overlapping spatial voxel set between processes. By calculating the ratio of the number of overlapping spatial voxels to the number of spatial voxels of a single process, the spatial overlap is obtained, thereby accurately quantifying the degree of intersection of processes in the spatial range and effectively reflecting the physical conflict between different processes in the work site.

[0152] At the time level, the minimum and maximum values ​​of the time dimension are determined from the four-dimensional voxel set of each process to obtain the process duration. Simultaneously, the time dimension range is extracted from the spatiotemporally overlapping voxel set to obtain the overlap duration. By comparing the overlap duration with the shortest of the two process durations, the time overlap degree is obtained, which reflects the degree to which the processes are synchronized on the time axis, effectively revealing potential conflicts in resource allocation and construction schedule.

[0153] Furthermore, this embodiment introduces a weighted mechanism when generating the final spatiotemporal conflict coefficient. Based on the varying sensitivities of spatial and temporal conflicts at the construction site, weights are assigned to spatial overlap and temporal overlap respectively. For example, in situations where the construction site is narrow and space is limited, the weight of spatial overlap can be appropriately increased to ensure that physical interference is resolved first; in scenarios with scarce resources and frequent overlapping of work processes, the weight of temporal overlap can be increased accordingly to avoid simultaneous occupation of manpower and equipment. Ultimately, the spatiotemporal conflict coefficient generated through a weighted summation method can comprehensively characterize the intensity of conflicts between work processes.

[0154] Please see Figure 4 The schematic diagram of spatiotemporal overlap voxels at time t in Embodiment 1 of the present invention, wherein X, Y, and Z are three-dimensional spatial coordinates, black circles are spatiotemporal overlap voxels, and gray circles are non-spatiotemporal overlap voxels.

[0155] The specific steps of S4.4 are as follows:

[0156] S4.4.1: Extract non-repeating spatial coordinates from the four-dimensional occupied voxel set to obtain the spatial occupied voxel set of the process, and extract non-repeating spatial coordinates from the spatiotemporal overlapping voxel set to obtain the overlapping spatial voxel set between processes.

[0157] In this embodiment, by preprocessing the four-dimensional occupancy voxel set, the four-dimensional data containing time and space attributes is reduced in dimensionality, retaining only the three-dimensional spatial coordinates (such as x, y, z) and removing duplicate coordinate points, thus obtaining the spatial occupancy voxel set of the process. This ensures that the spatial set can clearly reflect the physical operation scope involved in the process without being interfered with by redundant information in the time dimension.

[0158] Furthermore, when extracting the spatiotemporally overlapping voxel set, only spatial coordinate information is retained, and duplicate values ​​are removed, thus forming the overlapping spatial voxel set between processes. The overlapping spatial voxel set can intuitively represent the spatial overlap area of ​​two processes and is an important basis for calculating the degree of spatial overlap. For example, when process A (support shotcrete) and process B (reinforcement installation) are constructed simultaneously on the same tunnel cross section, the extracted overlapping spatial voxel set will be concentrated in the corresponding coordinate area, thus reflecting the actual situation of spatial intersection.

[0159] S4.4.2: Compare the spatial volumes between two processes to obtain the minimum value of the spatial volume between the two processes. Divide the overlapping spatial volume by the minimum value to obtain the spatial overlap degree. The spatial volume is the number of elements in the voxel set occupied by each process, and the overlapping spatial volume is the number of elements in the voxel set of the overlapping space between processes.

[0160] Specifically, during construction, simply calculating the area or volume of overlapping processes is often insufficient to accurately reflect the severity of conflicts. If the absolute overlap volume is used directly as the criterion, some overlaps in large-scale processes may be underestimated, while slight overlaps in small-scale processes may be exaggerated. Therefore, this embodiment uses a normalization strategy, that is, using the minimum of the spatial volumes of two processes as the denominator, to ratio the overlap volumes, thereby obtaining a spatial overlap index with relative significance.

[0161] In the specific implementation, firstly, unique coordinate points in the spatial dimension are extracted from the four-dimensional voxel set of the process, and the number of voxels is calculated as a measure of the process spatial volume. Then, unique spatial coordinates are extracted from the spatiotemporally overlapping voxel set to obtain the overlapping spatial voxel set between processes, and the number of elements is counted as the overlapping spatial volume. Finally, the measures of the spatial volume between processes are compared to determine the minimum spatial volume between two processes. Then, the overlapping spatial volume between two processes is divided by the minimum spatial volume between the two processes to obtain the spatial overlap index.

[0162] S4.4.3: Determine the minimum and maximum time values ​​of each process from the four-dimensional voxel set of the process, and calculate the process duration;

[0163] Specifically, in a construction scenario, the execution of each process involves not only the physical occupation of a spatial area but also a defined time interval, namely the entire process from start to finish. To accurately reflect the dynamic attributes of the process in the time dimension, this embodiment, after obtaining the four-dimensional voxel set of the process's occupation, analyzes the time coordinates within it to extract the minimum and maximum time values ​​involved in the process. The minimum time value corresponds to the start time of the process, and the maximum time value corresponds to the end time of the process; the difference between the two is the duration of the process.

[0164] S4.4.4: Determine the minimum and maximum time values ​​from the spatiotemporal overlap voxel set, calculate the overlap duration, compare the process durations between the two processes to obtain the shortest process duration between the two processes, and divide the overlap duration by the shortest value to obtain the time overlap degree;

[0165] Specifically, in the dynamic modeling of construction progress monitoring, simply identifying temporal overlap between processes is insufficient to accurately characterize the degree of conflict. To quantitatively measure the degree of synchronization of processes along the time axis, this embodiment extracts the minimum and maximum values ​​of the time dimension from the spatiotemporally overlapping voxel set, calculates the length of the coverage area as the overlap duration, and further normalizes it into the temporal overlap degree.

[0166] In the specific implementation process, the time coordinates of the overlapping voxel set are first analyzed to determine the earliest overlapping time point (minimum time value) and the latest overlapping time point (maximum time value). The difference between the minimum and maximum time values ​​is the length of the intersection interval of the processes in the time dimension, i.e., the overlapping duration. Subsequently, the process durations between processes are compared to determine the shortest process duration between two processes. Then, the overlapping duration describing the overlapping process of two processes is divided by the shortest process duration between the two processes to obtain the time overlap degree.

[0167] S4.4.5: Set weights based on spatial overlap and temporal overlap, and generate the spatiotemporal conflict coefficient between processes by weighted summation.

[0168] In this embodiment, spatial overlap and temporal overlap are combined, and a spatiotemporal conflict coefficient is generated through weighted fusion to achieve a comprehensive judgment of spatial and temporal dimensions. Specifically, in construction progress monitoring, relying solely on either spatial or temporal overlap is insufficient to fully reflect the conflicts between work processes. Spatial overlap primarily reveals the intersection of physical work areas, such as overlapping construction of two processes on the same tunnel section or bridge pier; while temporal overlap emphasizes the synchronicity of work progress, reflecting the overlap of different processes in their construction cycles. Considering only spatial overlap may underestimate the conflicts in equipment and personnel resources caused by simultaneous work processes; considering only temporal overlap may ignore physical interference caused by high spatial overlap.

[0169] In practical calculations, weighting coefficients are first assigned to spatial overlap and temporal overlap based on the management priorities of the construction site. For example, in confined work areas such as tunnels or shafts, limited space is often the primary challenge; in this case, the weight of spatial overlap can be increased to prioritize the identification and resolution of physical conflicts. Conversely, in high-intensity assembly line construction or resource-constrained scenarios, the temporal overlap between work processes has a greater impact on personnel and equipment scheduling; therefore, the weight of temporal overlap can be appropriately increased. This weighting adjustment mechanism allows for flexible adaptation to conflict sensitivities under different construction environments.

[0170] The weighted summation formula is as follows:

[0171]

[0172] in, Represents the spatiotemporal conflict coefficient. Indicates spatial overlap. Indicates the degree of time overlap. , Let be the weight parameters, and satisfy ? .

[0173] S4.5: Compare the spatiotemporal conflict coefficient with the preset conflict threshold to generate a conflict event library.

[0174] Specifically, in dynamic monitoring of construction progress, simply calculating the spatiotemporal conflict coefficient is insufficient for direct decision-making and scheduling. To achieve automatic identification and archiving of conflicts, this embodiment compares the calculated spatiotemporal conflict coefficient with a preset conflict threshold. When the conflict coefficient exceeds the threshold, the work process pair is marked as a conflict event and written into the conflict event database.

[0175] In practice, the conflict threshold can be configured according to the project scale, construction environment, and management requirements. For example, in large bridge or tunnel projects, due to the higher safety risks, the threshold can be set to a lower level to ensure sensitive detection of potential conflicts; while in general civil engineering projects, a relatively higher threshold can be set to reduce irrelevant alarms.

[0176] S4.5.1: When the spatiotemporal conflict coefficient is greater than or equal to the preset conflict threshold, the corresponding process pair is marked as a conflict event;

[0177] Specifically, in the dynamic monitoring of construction projects, the spatiotemporal conflict coefficient obtained by simple calculation is only a quantitative indicator. It still needs to be combined with a threshold judgment mechanism to be transformed into actionable conflict event information. This embodiment combines continuous numerical quantitative results with construction management decisions by setting a conflict threshold. When the spatiotemporal conflict coefficient of a work process pair is greater than or equal to the threshold, it is judged as a conflict event with practical management significance, and the work process pair is added to the conflict event database for resource scheduling optimization.

[0178] S4.5.2: When the spatiotemporal conflict coefficient is less than the preset conflict threshold, the process pair is marked as conflict-free;

[0179] Specifically, during the dynamic monitoring and optimization of construction progress, not all work process pairs present actual conflict risks. Recording and intervening in all work process pairs would not only waste resources but also potentially interfere with managers' accurate identification of critical conflicts. Therefore, this embodiment introduces a threshold mechanism in the conflict coefficient determination stage. When the spatiotemporal conflict coefficient of a work process pair is lower than a preset threshold, it is determined that there is no substantial interference in spatial location or temporal progress, thus marking the work process pair as conflict-free.

[0180] S4.5.3: Generate a conflict event library based on all conflict events.

[0181] Specifically, in the process of dynamic monitoring and optimization of construction progress, simply identifying conflicts in individual work process pairs is insufficient to support overall decision-making and scheduling. Therefore, this embodiment, after comparing the spatiotemporal conflict coefficients with thresholds, centrally organizes all work process pairs identified as conflicting, generating a structured conflict event database. As a global data container, the conflict event database can systematically store, retrieve, and analyze various conflict information, thereby providing a unified decision support platform for project management.

[0182] In the specific implementation process, key information for each conflict event is written into the conflict event database, including but not limited to: work process pair identifier, spatial overlap, temporal overlap, spatiotemporal conflict coefficient, conflict level, and the spatiotemporal range of the conflict. Through a unified data format and index structure, the conflict event database supports rapid cross-dimensional querying and comparison. For example, managers can query all associated conflicts by work process number, or retrieve conflict events occurring within a specific construction phase by time interval.

[0183] The specific steps for S5 are as follows:

[0184] S5.1: Obtain the corresponding process pairs and corresponding resource configuration data from the conflict event database, and optimize resource scheduling;

[0185] Specifically, in the digital progress monitoring system for construction projects, the conflict event database obtained during the conflict detection phase not only reflects the spatial and temporal conflict relationships between work processes but also provides key input for resource scheduling optimization. This embodiment analyzes each work process pair in the conflict event database to extract conflicting work process combinations and simultaneously acquires corresponding resource allocation data, such as personnel, equipment, material occupancy, and supply capacity.

[0186] S5.2: Establish a multi-objective optimization model with the objective functions of minimizing project delays, reducing resource idle rates, and avoiding spatial and temporal conflicts. The constraints of the multi-objective optimization model include process dependencies, resource supply limits, and construction safety distances.

[0187] Specifically, in the problem of optimizing construction schedule, a single objective function often fails to fully reflect the actual construction needs. For example, if the goal is only to minimize construction delays, some resources may be left idle for extended periods, reducing overall utilization efficiency; if the goal is simply to increase resource utilization, high-intensity overlapping operations may be introduced, leading to an increase in safety hazards and conflicts. Therefore, this embodiment establishes a multi-objective optimization model that, while ensuring the overall construction schedule, also considers resource allocation efficiency and construction safety.

[0188] Furthermore, this embodiment introduces a weighted coefficient and priority mechanism during the modeling process. This allows managers to dynamically adjust the weights of the three objective functions for different construction scenarios. For example, during a tight deadline phase, the weight of the schedule objective can be increased; during a resource-constrained phase, the weight of the resource utilization objective can be increased; and in high-risk operation scenarios, the safety objective is prioritized.

[0189] S5.3: Adjust resource configuration data using heuristic optimization algorithms to generate resource optimization schemes.

[0190] Specifically, after establishing the multi-objective optimization model, the key issue becomes how to efficiently solve it under complex constraints. While traditional exact algorithms (such as linear programming or integer programming) can guarantee the theoretical optimal solution, in large-scale construction scenarios, due to the large number of procedures and complex constraints, they often face problems of high computational complexity and slow convergence speed. Therefore, this embodiment introduces a heuristic optimization algorithm that iteratively approaches the optimal solution by simulating natural population behavior or evolutionary mechanisms, thereby obtaining a high-quality resource optimization solution within a limited time.

[0191] In the specific implementation process, the resource allocation scheme is first encoded as candidate solutions in the solution space. For example, parameters such as the adjustment strategy of each process on the time axis, the allocation ratio of equipment and manpower, and the measures to avoid resource conflicts are mapped as solution vectors. Subsequently, the heuristic algorithm iteratively updates the solution vectors. In each iteration, the fitness of the solutions is evaluated through the objective function, and the solutions with good performance are retained for the next round of search.

[0192] Taking the Particle Swarm Optimization (PSO) algorithm as an example, each candidate solution is regarded as a "particle," with its position representing a specific resource allocation scheme and its velocity representing the adjustment magnitude. During the iteration process, the particles dynamically adjust their own state based on the historical best solution and the global best solution of the swarm, thereby achieving swarm intelligent search.

[0193] Example 2

[0194] Please see Figure 5 This invention provides an embodiment: a project four-control digital monitoring and optimization system based on a dynamic model, the system comprising:

[0195] The data acquisition module is used to acquire planned progress data, real-time execution data, and resource configuration data;

[0196] The predictive analysis module is used to perform time-series predictive analysis on the real-time execution data, compare it with the planned progress data, calculate the process progress deviation, and generate a progress deviation event.

[0197] The dynamic model update module is used to maintain and update the dynamic schedule model, which includes process time attributes and spatial attributes.

[0198] The conflict detection module is used to input the schedule deviation events into the schedule dynamic model, update the schedule dynamic model, perform inter-process conflict detection, and generate a conflict event library.

[0199] The resource scheduling optimization module is used to optimize resource scheduling based on the conflict event database and through optimization algorithms to generate a resource optimization scheme.

[0200] The cost management module is used to realize the visual tracking and optimization control of costs throughout the entire construction life cycle, and to generate cost management reports;

[0201] In this embodiment, the cost management module is designed to enable visualized tracking and optimized control of costs throughout the entire lifecycle of a construction project. Its core function is to provide real-time cost management support by integrating multi-dimensional cost data from the project, combined with construction progress and dynamic models.

[0202] Specifically, the cost management module integrates multiple cost data points, such as labor hours (charged hourly), equipment rental (e.g., crane costs), material consumption (calculated based on MTO quantity and purchase price), and costs specific to a construction area (e.g., costs related to a particular work zone). This cost data is combined with corresponding items in the 3D model (e.g., pipes, steel structures) to achieve a triple mapping of "model-schedule-cost." Through this integration, every cost in the construction process can be closely linked to the schedule and construction area, ensuring dynamic updates and accurate tracking of cost data.

[0203] Furthermore, the cost management module overlays a heatmap layer onto the dynamic schedule model, displaying cost density by region or equipment type, allowing for a clear view of cost consumption for each region, equipment, and material. By comparing planned and actual cost trends using line charts, potential cost overruns can be identified in advance, providing timely warnings and assisting project managers in making adjustment decisions.

[0204] For example, taking a pipe unit in a 3D model (e.g., "MAT-304-001") as an example, its corresponding 4D task node (e.g., the "2024-Q3 Pipe Welding" task) is associated with it. By comparing the budget mapping (e.g., a labor budget of 700,000 yuan) with the actual cost data (e.g., actual labor cost of 798,000 yuan), the overspending rate (e.g., 14%) can be calculated. In this way, project managers can comprehensively track various costs, ensure real-time understanding of the cost status on the construction site, and promptly identify anomalies, avoiding project delays or resource waste caused by cost overspending.

[0205] The security management module is used to dynamically integrate the progress dynamic model with video surveillance to achieve video visualization management, and also integrate security reports with third-party systems.

[0206] In this embodiment, the core function of the safety management module is to achieve visualized safety management throughout the entire project lifecycle by dynamically integrating the 3D model with the video surveillance system. The design goal of this module is to improve the real-time performance and accuracy of on-site safety management, providing comprehensive support for project safety control.

[0207] Specifically, the safety management module can integrate third-party video surveillance data, supporting rapid management of data from each video surveillance point through the platform. All surveillance cameras and their related data at the project site can be centrally managed through this module, providing a unified view for monitoring the safety status of the construction site.

[0208] Furthermore, the safety management module dynamically binds the video surveillance system to the progress dynamic model, creating a correspondence between each monitoring point and its actual location on the construction site within the 3D model. This allows project managers to quickly view the location of each monitoring point in the 3D model and view the associated video data online in real time. This interconnectivity makes safety management more intuitive and efficient; users can not only view real-time video data from the site but also make more accurate safety judgments based on the actual 3D scene.

[0209] Furthermore, the safety management module supports integration with third-party safety reporting systems, importing safety report data into the platform. Safety reports can provide a summary and analysis of crucial information such as construction site safety inspections, accident records, and risk assessments, offering project managers real-time safety status feedback. By integrating safety reports, managers can ensure that project safety management complies with standards and make timely adjustments based on the data.

[0210] The quality monitoring module is used to perform visualized management of project quality throughout the entire project lifecycle, based on schedule deviation events, conflict event databases, resource optimization schemes, cost management reports, and safety reports.

[0211] In this embodiment, the quality monitoring module integrates a database of project schedule deviation events, conflict events, resource optimization plans, cost management reports, and safety reports to achieve full-cycle quality visualization management. This module monitors quality issues during construction in real time, identifying factors that may affect quality, such as schedule deviations or resource conflicts. For example, when a process is delayed or a time-space conflict occurs between processes, the system automatically issues a quality warning, helping managers adjust process arrangements or resource allocation in a timely manner to ensure that construction quality is not affected.

[0212] Furthermore, the quality monitoring module, through its integration with the safety management module, provides dual assurance for both quality and safety. By analyzing resource optimization plans and cost management data, the module ensures the rational allocation of resources, preventing quality issues caused by insufficient or improper resource allocation. Simultaneously, integrated safety reports help monitor potential safety hazards during construction, avoiding quality accidents resulting from safety problems. Through these comprehensive functions, the quality monitoring module provides comprehensive and real-time quality assurance for the project, ensuring timely and high-quality completion.

[0213] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention.

Claims

1. A project four-control digital monitoring optimization method based on a dynamic model, applied to the schedule, cost, safety, and quality management of engineering projects, characterized by: The method includes: Acquire planned progress data, real-time execution data, and resource allocation data for each process in an engineering project; Perform time-series predictive analysis on the real-time execution data, compare it with the planned progress data, calculate the process progress deviation, and generate a progress deviation event. The schedule deviation event is input into the schedule dynamic model to update the schedule dynamic model, wherein the schedule dynamic model is used to represent the dynamic state of the process in time and space. Based on the updated progress dynamic model, inter-process conflict detection is performed to generate a conflict event library. The inter-process conflict detection is based on the updated progress dynamic model to detect the spatiotemporal overlap between processes and to quickly calculate the spatiotemporal conflict coefficient through a sparse storage structure. Based on the conflict event database, resource scheduling is optimized using an optimization algorithm on the resource configuration data to generate a resource optimization scheme. The process of detecting inter-process conflicts and generating a conflict event library based on the updated progress dynamic model includes: The process nodes of the progress dynamic model are mapped to a four-dimensional coordinate system containing spatial and temporal attributes, generating a four-dimensional voxel set of the process. The four-dimensional occupied voxel set is written into a sparse storage structure, and a coarse screening is performed based on the four-dimensional boundary between processes to determine potentially overlapping process pairs. Based on the overlapping process pairs, voxel set intersection operation is performed through sparse storage structure to obtain spatiotemporally overlapping voxels between processes, and a spatiotemporally overlapping voxel set is generated. Based on the spatiotemporal overlap voxel set, the spatial overlap and temporal overlap between processes are calculated, and the spatiotemporal conflict coefficient between processes is generated. The spatiotemporal conflict coefficient is compared with a preset conflict threshold to generate a conflict event library; The step of calculating the spatial and temporal overlap between processes based on the spatiotemporal overlap voxel set, and generating the spatiotemporal conflict coefficient between processes, includes: The spatial coordinates of the process are extracted from the four-dimensional occupied voxel set to obtain the spatial occupied voxel set of the process, and the spatial coordinates of the process overlap are extracted from the spatiotemporal overlapping voxel set to obtain the overlapping spatial voxel set between processes. The spatial volume between two processes is compared to obtain the minimum value of the spatial volume between the two processes. The overlapping spatial volume is divided by the minimum value to obtain the spatial overlap. The spatial volume is the number of elements in the voxel set occupied by each process, and the overlapping spatial volume is the number of elements in the voxel set of the overlapping space between processes. Determine the minimum and maximum time values ​​of each process from the four-dimensional voxel set of the process, and calculate the process duration; Determine the minimum and maximum time values ​​from the spatiotemporal overlap voxel set, calculate the overlap duration, compare the process durations between two processes to obtain the shortest process duration between the two processes, and divide the overlap duration by the shortest value to obtain the time overlap degree. Weights are set based on spatial overlap and temporal overlap, and the spatiotemporal conflict coefficient between processes is generated by weighted summation.

2. The project four-control digital monitoring optimization method based on a dynamic model according to claim 1, characterized in that, Perform time-series predictive analysis on the real-time execution data, compare it with the planned schedule data, calculate the process schedule deviation, and generate schedule deviation events, including: The real-time execution data is preprocessed. The preprocessed real-time execution data is input into the time series prediction model to generate progress prediction data for future periods; The progress prediction data is compared with the planned progress data step by step to obtain the progress deviation data. Based on the schedule deviation data, the dependencies between processes are identified by combining the process dependency graph, the deviation-related processes are determined, and a set of deviation-related processes is obtained, wherein the process dependency graph is obtained by the critical path algorithm; Based on the schedule deviation data and the set of processes associated with the deviation, a schedule deviation event is generated.

3. The project four-control digital monitoring optimization method based on a dynamic model according to claim 2, characterized in that, The schedule forecast data is compared with the planned schedule data step by step to obtain schedule deviation data, including: The predicted completion time of the progress forecast data and the planned completion time of the planned progress data are obtained. The difference between the predicted completion time and the planned completion time of the process is calculated to obtain the progress deviation between the predicted progress data and the planned progress data. The schedule deviation is compared with a preset deviation threshold. When the schedule deviation exceeds the deviation threshold, the corresponding process is marked as having a schedule deviation, and the process with the schedule deviation is obtained. Based on the schedule deviation process and the corresponding schedule deviation amount, schedule deviation data is generated.

4. The project four-control digital monitoring optimization method based on a dynamic model according to claim 3, characterized in that, The schedule deviation event is input into the schedule dynamic model, and the schedule dynamic model is updated, including: The schedule deviation process and the deviation-related process in the schedule deviation event are mapped to the process nodes in the schedule dynamic model, wherein the process nodes include time attributes and space attributes; Update the time attributes of the process node based on the schedule deviation amount in the schedule deviation event, wherein the time attributes include start time, duration and end time; The progress dynamic model is updated based on the adjusted time and space attributes of the process nodes.

5. The project four-control digital monitoring optimization method based on a dynamic model according to claim 1, characterized in that, The spatiotemporal conflict coefficient is compared with a preset conflict threshold to generate a conflict event database, including: When the spatiotemporal conflict coefficient is greater than or equal to the preset conflict threshold, the corresponding process pair is marked as a conflict event; When the spatiotemporal conflict coefficient is less than the preset conflict threshold, the process pair is marked as conflict-free; A conflict event library is generated based on all conflict events.

6. The project four-control digital monitoring optimization method based on a dynamic model according to claim 1, characterized in that, Based on the conflict event database, resource scheduling is optimized using an optimization algorithm on the resource configuration data to generate a resource optimization scheme, including: Based on the conflict event database, obtain the corresponding work process pairs and corresponding resource configuration data, and optimize resource scheduling; A multi-objective optimization model is established with the objective functions of minimizing project delays and minimizing resource scheduling costs. The constraints of the multi-objective optimization model include process dependencies, resource supply limits, and construction safety distances. Resource optimization schemes are generated by adjusting resource configuration data using heuristic optimization algorithms.

7. The project four-control digital monitoring optimization method based on a dynamic model according to claim 6, characterized in that, The process involves adjusting resource allocation data using a heuristic optimization algorithm to generate a resource optimization scheme. The heuristic optimization algorithm is a particle swarm optimization algorithm, which includes: Initialize the particle swarm position and velocity, where the particle swarm position represents the resource optimization scheme and the velocity represents the adjustment range; The fitness of each particle is calculated based on the fitness function, which is a weighted average of project delay, resource idle rate, and number of conflict events. Update the individual particle's optimal position and the global optimal position; The particle swarm is iteratively updated based on the velocity update formula and the position update formula. When the number of iterations or fitness converges, the globally optimal resource optimization scheme is output.

8. A project four-control digital monitoring and optimization system based on a dynamic model, used to implement the project four-control digital monitoring and optimization method based on a dynamic model as described in any one of claims 1-7, characterized in that, The system includes: The data acquisition module is used to acquire planned progress data, real-time execution data, and resource configuration data; The predictive analysis module is used to perform time-series predictive analysis on the real-time execution data, compare it with the planned progress data, calculate the process progress deviation, and generate a progress deviation event. The dynamic model update module is used to maintain and update the dynamic schedule model, which includes process time attributes and spatial attributes. The conflict detection module is used to input the schedule deviation events into the schedule dynamic model, update the schedule dynamic model, perform inter-process conflict detection, and generate a conflict event library. The resource scheduling optimization module is used to optimize resource scheduling based on the conflict event database and through optimization algorithms to generate a resource optimization scheme. The cost management module is used to realize the visual tracking and optimization control of costs throughout the entire construction life cycle, and to generate cost management reports; The security management module is used to dynamically integrate the progress dynamic model with video surveillance to achieve video visualization management, and also integrate security reports with third-party systems. The quality monitoring module is used to perform visualized management of project quality throughout the entire project lifecycle, based on schedule deviation events, conflict event databases, resource optimization schemes, cost management reports, and safety reports.

Images