A key project construction progress and shutdown risk early warning method and system

Abstract

The application discloses a kind of key project construction progress and system and method for risk early warning, by obtaining the buffer capacity data of labor subsystem, mechanical subsystem, material subsystem, fund subsystem and management subsystem in construction system and the logical dependence relationship and planned construction period of each process in construction plan network diagram, based on the coupling strength data between each subsystem buffer capacity data and each subsystem, calculate system resilience entropy, based on the coupling strength between high influence process, the key degree of process, the number of redundant path, the total amount of float time and calculate topological elasticity coefficient;When system resilience entropy exceeds first preset threshold or topological elasticity coefficient is lower than second preset threshold, generate early warning signal;Further, in response to early warning signal, based on structural causal model, counterfactual reasoning is carried out to subsystem buffer capacity data, intervention scheme containing target subsystem, target buffer capacity adjustment amount and expected effect is generated and output, realize the automatic generation of early warning and intervention scheme.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, specifically to a method and system for early warning of construction progress and work stoppage risks in key projects. Background Technology

[0002] In the construction of large-scale key projects, managing the construction progress has always been a management challenge. Traditional progress management methods, such as critical chain technology, mainly determine whether the progress is lagging by monitoring the consumption of the process buffer, which is essentially a retrospective tracking of the "deviation between plan and reality." Other data mining-based methods provide early warnings by constructing construction maps and matching historical risk patterns, but their effectiveness is highly dependent on the completeness of historical data, and they often stop at "alarm" after the warning, failing to provide managers with specific response strategies. Although these methods have different focuses, they all focus on monitoring risk signals themselves, ignoring the inherent ability of the construction system as an organic whole composed of multiple subsystems such as labor, machinery, materials, and funds to resist disturbances and maintain progress. When the buffer capacity of a certain subsystem continues to decline, or the process network structure becomes abnormally fragile, the system is actually in a dangerous state of "collapse at the slightest touch," but existing technologies are unable to capture this systemic vulnerability before the risk event actually occurs. Summary of the Invention

[0003] In view of the above problems, the present invention provides a method and system for early warning of construction progress and work stoppage risks of key projects. By quantifying the overall ability of the construction system to resist disturbances and the vulnerability of the process network structure, and automatically generating intervention plans after the early warning, the invention achieves a leap from passive alarm to active prevention and decision support.

[0004] To achieve the above objectives, in a first aspect, this application provides a method for early warning of construction progress and work stoppage risks in key projects, including:

[0005] Obtain buffer capacity data for the labor subsystem, machinery subsystem, materials subsystem, funding subsystem, and management subsystem of the construction system;

[0006] Obtain the logical dependencies and planned durations of each process in the construction schedule network diagram;

[0007] Based on the buffer capacity data of each subsystem and the coupling strength data between each subsystem, the system resilience entropy is calculated.

[0008] Based on the criticality of each process, the number of redundant paths, the total floating time, and the coupling strength between high-impact processes in the construction plan network diagram, the topology elasticity coefficient is calculated.

[0009] When the system resilience entropy exceeds the first preset threshold or the topological elasticity coefficient is lower than the second preset threshold, an early warning signal is generated.

[0010] In response to the early warning signal, counterfactual reasoning is performed on the subsystem buffer capacity data based on the structural causal model to generate at least one intervention plan. The intervention plan includes the target subsystem, the target buffer capacity adjustment amount, and the expected effect.

[0011] Output the intervention plan and expected results.

[0012] Furthermore, based on the buffering capacity data of each subsystem and the coupling strength data between each subsystem, the system resilience entropy is calculated, including:

[0013] The buffer capacity data of the labor subsystem, machinery subsystem, materials subsystem, funds subsystem, and management subsystem are normalized respectively to generate normalized buffer capacity values ​​for each subsystem.

[0014] Based on the transmission time interval between fund delay payment events and material supplier supply disruption events in historical event data, the first coupling strength value between the fund subsystem and the material subsystem is determined.

[0015] Based on the transmission time interval between fund delay payment events and subcontractor passive work events in historical event data, the second coupling strength value between the fund subsystem and the labor subsystem is determined.

[0016] Based on the transmission time interval between material shortage events and mechanical downtime events in historical event data, the third coupling strength value between the material subsystem and the mechanical subsystem is determined.

[0017] Based on the transmission time interval between worker absence events and machine operator shutdown events in historical event data, the fourth coupling strength value between the labor subsystem and the machinery subsystem is determined.

[0018] Based on the set of transmission time intervals between approval delay events and response delay events of each subsystem in historical event data, the fifth coupling strength value between the management subsystem and each other subsystem is determined.

[0019] The vulnerability coefficient of each subsystem is generated by taking the reciprocal of each normalized buffer capacity value.

[0020] The system resilience entropy is generated by weighted summation of each vulnerability coefficient and its corresponding coupling strength value.

[0021] Furthermore, based on the transmission time interval between fund delay payment events and material supplier supply disruption events in historical event data, a first coupling strength value between the fund subsystem and the material subsystem is determined, including:

[0022] Extract the timestamp sequences of fund delay payment events and material supplier supply disruption events from historical event data;

[0023] Time window matching is performed on the timestamp sequence of fund delay payment events and the timestamp sequence of material supplier supply interruption events to determine the material supplier supply interruption events that occur within the first preset time window after each fund delay payment event, as related event pairs.

[0024] Calculate the time difference between the delayed payment event and the material supplier supply interruption event in each related event pair to generate a propagation time difference sequence;

[0025] Outlier removal is performed on the conduction time difference sequence, and the conduction time difference within the second preset information interval is retained to generate a valid conduction time difference sequence.

[0026] Calculate the mean of the effective conduction time difference sequence to generate the average conduction time difference;

[0027] The reciprocal of the average conduction time difference is used as the first coupling strength value.

[0028] Furthermore, the vulnerability coefficients are weighted and summed with their corresponding coupling strength values ​​to generate the system resilience entropy, including:

[0029] Acquire historical shutdown event data, and extract the time-series change rate of each vulnerability coefficient and the time-series change rate of each coupling strength value before each historical shutdown event occurs from the historical shutdown event data;

[0030] The time-series rate of change of vulnerability coefficients and the time-series rate of change of coupling strength values ​​are normalized to generate normalized values ​​of each vulnerability rate of change and each coupling strength rate of change.

[0031] The joint sensitivity coefficient of each subsystem pair is generated by multiplying the normalized value of the change rate of each vulnerability with the normalized value of the change rate of the corresponding coupling strength.

[0032] The joint sensitivity coefficients are normalized to generate dynamic weights for each subsystem pair;

[0033] Each vulnerability coefficient is multiplied by its corresponding coupling strength value to generate the vulnerability-coupling product of each subsystem pair;

[0034] Multiply each fragile-coupling product by its corresponding dynamic weight to generate a weighted contribution value for each subsystem pair;

[0035] Sum all weighted contribution values ​​to generate the system resilience entropy.

[0036] Furthermore, based on the criticality of each process, the number of redundant paths, the total float time, and the coupling strength between high-impact processes in the construction plan network diagram, the topology resilience coefficients are calculated, including:

[0037] Monte Carlo simulation was performed on each process in the construction plan network diagram. The duration of each process was randomly disturbed, and the average impact of the duration disturbance of each process on the total duration was calculated. The average impact was used as the criticality of the process.

[0038] The number of paths containing processes with a criticality lower than a preset criticality threshold in the construction plan network diagram is counted as the number of redundant paths.

[0039] The sum of the floating times of all processes in the construction schedule network diagram is used as the total floating time.

[0040] For processes with a criticality higher than a preset criticality threshold, pair them up in pairs, calculate the immediate predecessor and immediate successor time interval in the logical dependency relationship between each paired process, and use the reciprocal of the immediate predecessor and immediate successor time interval as the coupling strength between the paired processes.

[0041] The maximum value among all coupling strengths is selected as the coupling strength between high-impact processes;

[0042] Add the number of redundant paths to the total floating time to generate the elastic gain term;

[0043] The mean of criticality is multiplied by the coupling strength between high-impact processes to generate an elastic loss term;

[0044] The ratio of the elastic gain term to the elastic loss term is used as the topological elasticity coefficient.

[0045] Furthermore, Monte Carlo simulations were performed on each process in the construction schedule network diagram, randomly perturbing the duration of each process. The average impact of the duration perturbation on the total duration was statistically analyzed, and this average impact was used as the criticality of the process, including:

[0046] For each process in the construction plan network diagram, a sequence of random schedule disturbance values ​​that follows a preset probability distribution is generated. The preset probability distribution is calibrated based on the historical schedule deviation data of the process.

[0047] Each disturbance value in the random duration disturbance value sequence is sequentially superimposed onto the planned duration of the corresponding process to generate a set of disturbance-post duration values;

[0048] Substitute the disturbed project duration value into the construction schedule network diagram, recalculate the critical path along the logical dependencies in the construction schedule network diagram, and generate the disturbed total project duration corresponding to each disturbance value.

[0049] Calculate the difference between the total project duration after the disturbance and the original planned total project duration, and generate the project duration deviation corresponding to each disturbance value;

[0050] The average value of all schedule deviations is calculated to generate the average impact of each process on the total schedule, which is used as the criticality of the process.

[0051] Furthermore, the number of redundant paths is added to the total floating time to generate a resilience gain term; the mean criticality is multiplied by the coupling strength between high-impact processes to generate a resilience loss term; the ratio of the resilience gain term to the resilience loss term is used as the topology resilience coefficient, including:

[0052] Calculate the criticality of all processes in the construction plan network diagram, calculate the arithmetic mean of all criticalities, and generate the average criticality.

[0053] Multiply the average criticality by the coupling strength between high-impact processes to generate an elastic loss term;

[0054] Calculate the floating time of all processes in the construction plan network diagram, sum the floating time of each process, and generate the total floating time.

[0055] The number of paths containing processes with a criticality lower than a preset criticality threshold in the construction plan network diagram is counted as the number of redundant paths.

[0056] Add the number of redundant paths to the total floating time to generate the elastic gain term;

[0057] Calculate the ratio of the elastic gain term to the elastic loss term, and use the ratio as the topological elasticity coefficient.

[0058] Furthermore, when the system resilience entropy exceeds a first preset threshold or the topological elasticity coefficient falls below a second preset threshold, a warning signal is generated, including:

[0059] Acquire historical shutdown event data, extract the system resilience entropy value corresponding to the time of occurrence of each historical shutdown event from the historical shutdown event data, and generate a historical resilience entropy sequence;

[0060] Extract the topological elasticity coefficient value corresponding to the time of occurrence of each historical shutdown event from the historical shutdown event data, and generate a historical elasticity coefficient sequence;

[0061] Perform percentile statistics on the historical resilience entropy sequence, and use the resilience entropy value corresponding to the Nth percentile as the first preset threshold.

[0062] Perform percentile statistics on the historical elasticity coefficient sequence, and use the elasticity coefficient value corresponding to the Mth percentile as the second preset threshold.

[0063] Real-time monitoring of the system resilience entropy at the current moment to determine whether the system resilience entropy is greater than the first preset threshold;

[0064] If the system resilience entropy is greater than the first preset threshold, a first warning signal is generated, which indicates the risk of shutdown due to insufficient system resilience.

[0065] Real-time monitoring of the topology elasticity coefficient at the current moment to determine whether the topology elasticity coefficient is less than the second preset threshold;

[0066] If the topology resilience coefficient is less than the second preset threshold, a second early warning signal is generated, which indicates the risk of work stoppage caused by the fragility of the progress network.

[0067] If the system resilience entropy is greater than the first preset threshold and the topology elasticity coefficient is less than the second preset threshold, a third warning signal is generated. The third warning signal indicates the risk of work stoppage caused by the superposition of insufficient system resilience and vulnerability of the schedule network.

[0068] Furthermore, percentile statistics are performed on the historical resilience entropy sequence, and the resilience entropy value corresponding to the Nth percentile is used as the first preset threshold, including:

[0069] Obtain all resilience entropy values ​​from the historical resilience entropy sequence, sort all resilience entropy values ​​in ascending order, and generate an ordered resilience entropy sequence.

[0070] Calculate the length of the ordered resilience entropy sequence, multiply the length by the preset percentile value, and generate the index position value;

[0071] Round the index value down to generate an integer index;

[0072] Extract the resilience entropy value at the integer index position from the ordered resilience entropy sequence as a candidate threshold;

[0073] Calculate the false alarm rate and false negative rate corresponding to the candidate threshold. The false alarm rate is the probability that the system resilience entropy exceeds the candidate threshold when no shutdown event occurs, and the false negative rate is the probability that the system resilience entropy does not exceed the candidate threshold when a shutdown event occurs.

[0074] The candidate threshold corresponding to the minimum sum of false positive rate and false negative rate is taken as the first preset threshold.

[0075] In a second aspect, the present invention also provides a key project construction progress and work stoppage risk early warning system, applicable to the method described in the first aspect. The system includes a data acquisition module, a resilience entropy calculation module, an elasticity coefficient calculation module, an early warning generation module, a counterfactual reasoning module, and an output module. The data acquisition module is used to acquire buffer capacity data for the labor subsystem, machinery subsystem, material subsystem, funding subsystem, and management subsystem of the construction system, as well as the logical dependencies and planned durations of each process in the construction plan network diagram. The resilience entropy calculation module is connected to the data acquisition module and is used to calculate the system resilience entropy based on the buffer capacity data of each subsystem and the coupling strength data between the subsystems. The elasticity coefficient calculation module... The module is connected to the data acquisition module and is used to calculate the topological elasticity coefficient based on the criticality of each process, the number of redundant paths, the total floating time, and the coupling strength between high-impact processes in the construction plan network diagram. The early warning generation module is connected to the resilience entropy calculation module and the elasticity coefficient calculation module respectively and is used to generate an early warning signal when the system resilience entropy exceeds a first preset threshold or the topological elasticity coefficient is lower than a second preset threshold. The counterfactual reasoning module is connected to the early warning generation module and is used to generate at least one intervention plan based on the structural causal model of the subsystem buffer capacity data in response to the early warning signal. The intervention plan includes the target subsystem, the target buffer capacity adjustment amount, and the expected effect. The output module is connected to the counterfactual reasoning module and is used to output the intervention plan and the expected effect.

[0076] Unlike existing technologies, the above-mentioned technical solution provides a method and system for early warning of construction progress and work stoppage risks in key projects. It acquires buffer capacity data from the labor, machinery, materials, funding, and management subsystems within the construction system, along with the logical dependencies and planned durations of each process in the construction plan network diagram. Based on the buffer capacity data of each subsystem and the coupling strength data between them, it calculates the system resilience entropy. It then calculates the topological elasticity coefficient based on the criticality of each process, the number of redundant paths, the total floating time, and the coupling strength between high-impact processes. When the system resilience entropy exceeds a first preset threshold or the topological elasticity coefficient falls below a second preset threshold, an early warning signal is generated. In response to the early warning signal, counterfactual reasoning is performed on the subsystem buffer capacity data based on a structural causal model to generate and output an intervention plan that includes the target subsystem, the target buffer capacity adjustment amount, and the expected effect. This achieves automatic generation of advanced early warning and intervention plans.

[0077] The above description of the invention is merely an overview of the technical solution of this application. In order to enable those skilled in the art to better understand the technical solution of this application and to implement it based on the description and drawings, and to make the above-mentioned objectives and other objectives, features and advantages of this application easier to understand, the following description is provided in conjunction with the specific embodiments and drawings of this application. Attached Figure Description

[0078] The accompanying drawings are only used to illustrate the principles, implementation methods, applications, features, and effects of specific embodiments of the present invention and other related contents, and should not be considered as limitations on this application.

[0079] In the accompanying drawings of the instruction manual:

[0080] Figure 1 This is a schematic diagram illustrating steps S101 to S107 of the method described in the specific implementation embodiment;

[0081] Figure 2 This is a schematic diagram illustrating steps S201 to S208 of the method described in a specific implementation.

[0082] Figure 3 This is a schematic diagram illustrating steps S301 to S308 of the method described in a specific implementation.

[0083] Figure 4 This is a schematic diagram illustrating steps S401 to S406 of the method described in a specific embodiment;

[0084] Figure 5 This is a schematic diagram of the early warning system described in a specific implementation.

[0085] The reference numerals used in the above figures are explained as follows:

[0086] 1. Early warning system; 11. Data acquisition module; 12. Resilience entropy calculation module; 13. Elasticity coefficient calculation module; 14. Early warning generation module; 15. Counterfactual reasoning module; 16. Output module. Detailed Implementation

[0087] To illustrate the possible application scenarios, technical principles, implementable specific solutions, and achievable objectives and effects of this application in detail, the following description, in conjunction with the listed specific embodiments and accompanying drawings, provides a detailed explanation. The embodiments described herein are merely illustrative of the technical solutions of this application and are therefore intended to limit the scope of protection of this application.

[0088] In this document, the term "embodiment" means that a specific feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The term "embodiment" appearing in various places throughout the specification does not necessarily refer to the same embodiment, nor does it specifically limit its independence or connection with other embodiments. In principle, in this application, as long as there are no technical contradictions or conflicts, the technical features mentioned in each embodiment can be combined in any way to form corresponding implementable technical solutions.

[0089] Unless otherwise defined, the technical terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains; the use of related terms herein is merely for the purpose of describing particular embodiments and is not intended to limit this application.

[0090] In the description of this application, the term "and / or" is used to describe the logical relationship between objects, indicating that three relationships can exist. For example, A and / or B means: A exists, B exists, and A and B exist simultaneously. Additionally, the character " / " in this document generally indicates that the preceding and following objects have an "or" logical relationship.

[0091] In this application, terms such as “first” and “second” are used only to distinguish one entity or operation from another, and do not necessarily require or imply any actual quantity, hierarchy or order relationship between these entities or operations.

[0092] Without further limitations, the use of terms such as “comprising,” “including,” “having,” or other similar open-ended expressions in this application is intended to cover non-exclusive inclusion, which does not exclude the presence of additional elements in a process, method, or product that includes the stated elements, such that a process, method, or product that includes a list of elements may include not only those defined elements but also other elements not expressly listed, or elements inherent to such a process, method, or product.

[0093] As understood in the Examination Guidelines, in this application, expressions such as "greater than," "less than," and "exceeding" are understood to exclude the stated number; expressions such as "above," "below," and "within" are understood to include the stated number. Furthermore, in the description of the embodiments in this application, "multiple" means two or more (including two), and similar expressions related to "multiple" are also understood in this way, such as "multiple groups" and "multiple times," unless otherwise explicitly specified.

[0094] In the description of the embodiments of this application, the space-related expressions used, such as "center," "longitudinal," "lateral," "length," "width," "thickness," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "vertical," "top," "bottom," "inner," "outer," "clockwise," "counterclockwise," "axial," "radial," and "circumferential," indicate the orientation or positional relationship based on the orientation or positional relationship shown in the specific embodiments or drawings. They are only for the purpose of describing the specific embodiments of this application or for the reader's understanding, and do not indicate or imply that the device or component referred to must have a specific position, a specific orientation, or be constructed or operated in a specific orientation. Therefore, they should not be construed as limitations on the embodiments of this application.

[0095] The processor described in the embodiments of this application can be implemented by hardware, firmware, software, or a combination thereof. It can be a circuit, one or more of an application-specific integrated circuit (ASIC), a digital signal processor (DSP), a digital signal processing device (DSPD), a programmable logic device (PLD), a field-programmable gate array (FPGA), a central processing unit (CPU), a controller, a microcontroller, or a microprocessor. It also includes other physical, biological, or chemical structures that can implement the same or equivalent functions as the processors listed above, such as biological neurons, quantum computing units, DNA computing units, etc., so that the processor can execute some or all of the steps in the computer program or method involved in the various embodiments of this application, or any combination of the steps mentioned therein.

[0096] The computer program involved in the embodiments can be stored in a computer device readable storage medium, which includes, but is not limited to, disks, magnetic tapes, magnetic cards, floppy disks, flash memory, optical disks, optical cards, read-only memory (ROM), random access memory (RAM), erasable programmable ROM (EPROM), and electrically erasable programmable ROM (EEPROM), etc., and also includes other biological, physical, or chemical structures that can achieve the same or equivalent functions as the storage media listed above, such as DNA, RNA, proteins, and other units with information storage capabilities. In specific embodiments, the storage medium involved can be one of the above-mentioned media types, or a combination of the above-mentioned media types. In different embodiments, the computer program involved in the embodiments can be centrally stored in a single medium, or distributed and stored in multiple media. The memory containing the computer device readable storage medium can be non-volatile memory or random access memory. These computer device readable storage media can be built into the device, or can be connected to the device involved in the embodiments as an external device or part of an external device. In some embodiments, the memory having a computer device readable storage medium is deployed locally; in other embodiments, the memory may be deployed remotely from the processor, for example, as a network-attached memory accessed via RF circuitry or an external port and a communication network, wherein the communication network may be the Internet, one or more intranets, a local area network (LAN), a wide area network (WLAN), a storage area network (SAN), or a suitable combination thereof, as long as computer device access to the memory is enabled. Furthermore, the computer program involved in the embodiments may be stored in plaintext / ciphertext form, or it may be designed as training data, integrated and recombined through model training and implicitly stored in the parameter states of a deep neural network or other machine learning model.

[0097] Please see Figure 1 In a first aspect, this embodiment provides a method for early warning of construction progress and work stoppage risks in key projects, including:

[0098] S101. Obtain buffer capacity data for the labor subsystem, machinery subsystem, material subsystem, funding subsystem, and management subsystem of the construction system.

[0099] S102. Obtain the logical dependencies and planned durations of each process in the construction plan network diagram;

[0100] S103. Based on the buffer capacity data of each subsystem and the coupling strength data between each subsystem, the system resilience entropy is calculated.

[0101] S104. Based on the criticality of each process, the number of redundant paths, the total floating time, and the coupling strength between high-impact processes in the construction plan network diagram, the topology elasticity coefficient is calculated.

[0102] S105. When the system resilience entropy exceeds the first preset threshold or the topological elasticity coefficient is lower than the second preset threshold, an early warning signal is generated.

[0103] S106. In response to the warning signal, perform counterfactual reasoning on the subsystem buffer capacity data based on the structural causal model to generate at least one intervention plan. The intervention plan includes the target subsystem, the target buffer capacity adjustment amount, and the expected effect.

[0104] S107. Output the intervention plan and expected results.

[0105] In step S101, the buffer capacity data of the labor subsystem characterizes the redundancy reserve level of the construction workforce to maintain construction progress when faced with staff absences or turnover; the buffer capacity data of the machinery subsystem reflects the backup capacity of construction machinery and equipment when encountering malfunctions or scheduling conflicts; the buffer capacity data of the materials subsystem reflects the safety reserve depth of material inventory in response to supply chain disruptions or transportation delays; the buffer capacity data of the financial subsystem measures the cash flow elasticity of the project when payment is delayed or costs exceed budget; and the buffer capacity data of the management subsystem characterizes the decision-making and response speed of the project management team in the face of emergencies. The buffer capacity data of each subsystem can be collected in real time or imported periodically through the labor management module, equipment management module, material management module, financial management module, and approval management module of the construction management information system, providing quantitative input for assessing the ability of the construction system to resist disturbances in various dimensions.

[0106] In step S102, the construction schedule network diagram is a directed graph structure with work processes as nodes and logical dependencies between work processes as edges. Logical dependencies include time-series constraints such as finish-start and start-start. The planned duration is the estimated duration of each work process under normal construction conditions. This information is exported from the building information model or project management software and forms the baseline framework for schedule management.

[0107] In step S103, the buffer capacity data of each subsystem is processed to characterize the reserve level of each subsystem in response to disturbances. The coupling strength data between subsystems is determined by analyzing the propagation speed of disturbances between different subsystems in historical events; higher coupling strength indicates faster disturbance propagation. System resilience entropy is used to quantify the overall vulnerability of the construction system. When the buffer capacity of each subsystem decreases and the coupling between subsystems is tight, the system resilience entropy increases, indicating that the system's ability to resist disturbances is weakened, providing a basis for early warning.

[0108] In step S104, the criticality of a process reflects the impact of its delay on the overall project duration; the number of redundant paths represents the number of alternative construction routes in the construction plan; the total float time represents the total time reserve for each process to be delayed without affecting the overall project duration; and the coupling strength between high-impact processes reflects the tightness of the logical dependency between processes with high criticality. The topological resilience coefficient characterizes the structural robustness of the construction schedule network. The number of redundant paths and the total float time constitute the resilience factor of the schedule network, while the number of high-impact processes and their tight coupling constitute the vulnerability factor. The ratio of the resilience factor to the vulnerability factor is used as the topological resilience coefficient. A higher resilience coefficient indicates a higher number of redundant paths and float time, while a lower coefficient indicates a higher number of high-impact processes that are tightly coupled. A higher topological resilience coefficient indicates that the schedule network can better absorb process delays, while a lower coefficient indicates a more vulnerable network.

[0109] In step S105, the first preset threshold is a pre-calibrated risk warning value for the system resilience entropy, which can be determined based on the distribution of system resilience entropy before a work stoppage event in historical construction data; the second preset threshold is a pre-calibrated risk warning value for the topological elasticity coefficient, which can be determined based on the distribution of topological elasticity coefficient before a work stoppage event in historical construction data. When the system resilience entropy exceeds the first preset threshold, it indicates that the overall resistance to disturbances of the construction system has decreased to a dangerous level; when the topological elasticity coefficient is lower than the second preset threshold, it indicates that the structural robustness of the construction schedule network is insufficient. A warning signal is automatically triggered when either of the above conditions is met, converting the risk state into an identifiable management signal. This dual-threshold triggering mechanism enables parallel monitoring of insufficient system resilience and the vulnerability of the schedule network.

[0110] In step S106, the structural causal model is constructed based on the causal relationship between the buffer capacity of each subsystem and the shutdown event in historical project data, describing the causal effect path between variables through a directed graph structure. Counterfactual reasoning, within the framework of the structural causal model, simulates the adjustment of the buffer capacity level of a specific subsystem through hypothetical intervention operations, predicting the changing trend of the system resilience entropy or topological elasticity coefficient after the intervention. In the generated intervention plan, the target subsystem is the subsystem that contributes the most to the system resilience entropy. The target buffer capacity adjustment amount is calculated based on the minimum adjustment required to restore the system resilience entropy or topological elasticity coefficient to a safe threshold. The expected effect is presented as the predicted value of the system resilience entropy or topological elasticity coefficient after the intervention.

[0111] In step S107, the intervention plan and expected results are output in a structured form to the project management platform or decision support interface. The output information may include the name of the target subsystem, the specific measures to be adjusted, the adjustment range, the expected effect value, and the intervention priority ranking, for managers to refer to and implement.

[0112] This embodiment constructs two evaluation indicators—system resilience entropy and topological elasticity coefficient—by collecting five dimensions of disturbance resistance reserve data from the construction system and structured information from the progress network. These indicators quantify the construction system's risk resistance capability from two dimensions: the overall disturbance resistance capability of the construction system and the structural robustness of the progress network. Based on this, an early warning is achieved through a dual-threshold triggering mechanism. Combined with counterfactual reasoning from a structural causal model, an intervention plan containing specific adjustment objects, adjustment ranges, and expected effects is automatically generated after the warning is triggered. This provides managers with an integrated solution from risk identification to decision support, enabling comprehensive management of the construction progress and shutdown risks of key projects.

[0113] Please see Figure 2 In some embodiments, the system resilience entropy is calculated based on the buffer capacity data of each subsystem and the coupling strength data between the subsystems, including:

[0114] S201. Normalize the buffer capacity data of the labor subsystem, machinery subsystem, material subsystem, capital subsystem, and management subsystem respectively to generate normalized buffer capacity values ​​for each subsystem.

[0115] S202. Based on the transmission time interval between fund delay payment events and material supplier supply interruption events in historical event data, determine the first coupling strength value between the fund subsystem and the material subsystem;

[0116] S203. Based on the transmission time interval between fund delay payment events and subcontractor passive work events in historical event data, determine the second coupling strength value between the fund subsystem and the labor subsystem;

[0117] S204. Based on the transmission time interval between material shortage events and mechanical downtime events in historical event data, determine the third coupling strength value between the material subsystem and the mechanical subsystem;

[0118] S205. Based on the transmission time interval between worker absence events and machine operator shutdown events in historical event data, determine the fourth coupling strength value between the labor subsystem and the machinery subsystem;

[0119] S206. Based on the set of transmission time intervals between approval delay events and response delay events of each subsystem in historical event data, determine the fifth coupling strength value between the management subsystem and each other subsystem.

[0120] S207. Take the reciprocal of each normalized buffer capacity value to generate the vulnerability coefficient of each subsystem.

[0121] S208. The system resilience entropy is generated by weighted summation of each vulnerability coefficient and its corresponding coupling strength value.

[0122] In step S201, normalization is used to transform the buffer capacity data of each subsystem to a unified numerical range, thereby eliminating the scale bias caused by different units of measurement of each indicator (such as manpower redundancy as a percentage, safety stock days as days, and decision response speed as hours) to subsequent comprehensive calculations. Specifically, the minimum-maximum normalization method can be used to linearly map the buffer capacity data of each subsystem to the interval between 0 and 1, generating normalized buffer capacity values ​​for each subsystem, providing standardized input for subsequent cross-dimensional comparisons and aggregations.

[0123] In step S202, the first coupling strength value is used to quantify the speed and probability of disturbances in the funding subsystem propagating to the materials subsystem. Historical event data is extracted from the event logs of the project management system or collaborative records between the finance and materials departments. Delayed payment events can be identified by the actual payment date of accounts payable in the finance system being later than the contractually agreed payment date, while supplier supply disruption events can be identified by interruptions or delays in supplier delivery records in the materials management system. The shorter the transmission time interval between delayed payment events and supplier supply disruption events, the more easily funding problems can trigger material supply disruptions, and the higher the first coupling strength value.

[0124] In step S203, the second coupling strength value is used to quantify the speed and probability of disturbances in the funding subsystem being transmitted to the labor subsystem. Specifically, subcontractor apathy events are determined based on whether the proportion of actual attendance on a given day that is lower than the planned attendance exceeds a preset threshold, or whether the proportion of completed work on a given day that is lower than the planned work exceeds a preset threshold. The shorter the transmission time interval between delayed payment events and subcontractor apathy events, the more easily the delay in funding can quickly lead to a decline in the work efficiency of the labor force, and the higher the second coupling strength value.

[0125] In step S204, the third coupling strength value is used to quantify the speed and probability of disturbances in the material subsystem being transmitted to the mechanical subsystem. Material shortage events are determined based on the inventory level recorded in the materials management system being lower than a preset multiple of the safety stock days, while mechanical downtime events are determined based on the unplanned downtime caused by lack of available materials recorded in the equipment management system. The shorter the transmission time interval between material shortage events and mechanical downtime events, the more easily material shortages can lead to rapid mechanical downtime, and the higher the third coupling strength value.

[0126] In step S205, the fourth coupling strength value is used to quantify the speed and probability of disturbances from the labor subsystem propagating to the mechanical subsystem. Worker absence events can be identified by the actual attendance being lower than the planned attendance in the labor management system's attendance records. Machine operator downtime events are determined based on the unplanned downtime duration recorded in the equipment management system due to operator absence. The shorter the propagation time interval between worker absence events and machine operator downtime events, the more quickly worker absences can lead to machine operator downtime, and the higher the fourth coupling strength value.

[0127] In step S206, the fifth coupling strength value is used to quantify the speed and probability of disturbances from the management subsystem propagating to other subsystems. Approval delay events are judged based on the deviation between the actual completion time and the planned completion time of the approval process recorded in the project approval system. Response delay events for each subsystem are judged based on the actual response time of the corresponding subsystem after approval exceeding a preset response time limit. Since the decision-making and approval process of the management subsystem involves multiple stages such as labor allocation approval, material procurement approval, and fund payment approval, its disturbances may simultaneously affect multiple subsystems. Therefore, the fifth coupling strength value is a comprehensive coupling metric.

[0128] In step S207, a higher normalized buffer capacity value indicates that the subsystem has a more sufficient reserve to resist disturbances, and its reciprocal, the vulnerability coefficient, is lower, indicating that the subsystem is not easily affected by disturbances; conversely, a lower normalized buffer capacity value and a higher vulnerability coefficient indicate that the subsystem is more vulnerable to disturbances.

[0129] In step S208, the vulnerability coefficient of each subsystem is multiplied by the coupling strength value of each subsystem pair that the subsystem participates in, and then all products are summed to obtain the system resilience entropy. The higher this value, the more vulnerable the construction system as a whole is, and the worse its ability to resist disturbances.

[0130] This embodiment eliminates the dimensional differences in the buffer capacity data of each subsystem through normalization processing. Based on the transmission time interval between specific disturbance event pairs in historical event data, it determines the coupling strength values ​​of five key subsystem pairs. The buffer capacity is then converted into a vulnerability coefficient by taking its reciprocal. Finally, the vulnerability coefficient and the coupling strength value are weighted and summed to achieve the quantitative calculation of the system's resilience entropy. This embodiment organically integrates the disturbance resistance reserve levels of various dimensions of the construction system with the disturbance transmission characteristics between subsystems, providing a quantitative basis for subsequent early warning triggering.

[0131] In some embodiments, determining a first coupling strength value between the funding subsystem and the materials subsystem based on the transmission time interval between fund delay payment events and material supplier supply disruption events in historical event data includes:

[0132] Extract the timestamp sequences of fund delay payment events and material supplier supply disruption events from historical event data;

[0133] Time window matching is performed on the timestamp sequence of fund delay payment events and the timestamp sequence of material supplier supply interruption events to determine the material supplier supply interruption events that occur within the first preset time window after each fund delay payment event, as related event pairs.

[0134] Calculate the time difference between the delayed payment event and the material supplier supply interruption event in each related event pair to generate a propagation time difference sequence;

[0135] Outlier removal is performed on the conduction time difference sequence, and the conduction time difference within the second preset information interval is retained to generate a valid conduction time difference sequence.

[0136] Calculate the mean of the effective conduction time difference sequence to generate the average conduction time difference;

[0137] The reciprocal of the average conduction time difference is used as the first coupling strength value.

[0138] In this embodiment, the timestamp sequence of delayed payment events can be extracted from the accounts payable module of the financial system. The actual payment date of each account payable is later than the contractually agreed payment date, and the specific date of each delayed payment event is recorded. Similarly, the timestamp sequence of supplier supply disruption events can be extracted from the supplier delivery records of the materials management system. The supplier's failure to deliver goods for consecutive periods exceeding the agreed delivery cycle is used as the criterion, and the specific date of each supply disruption event is recorded. Both timestamp sequences are arranged chronologically to form a structured event time index.

[0139] When performing time window matching between the timestamp sequences of delayed payment events and material supplier supply disruption events, the occurrence time of each delayed payment event is used as a reference point, and the system scans backward within a first preset time window to check for the occurrence of a material supplier supply disruption event. The length of the first preset time window is determined based on the average interval distribution of the two types of events in historical data; for example, 90% of the historical event interval is used as the window length to ensure coverage of the vast majority of causally related event pairs. If multiple material supplier supply disruption events occur within the first preset time window after a delayed payment event, the earliest occurring material supplier supply disruption event is paired with the delayed payment event as a related event pair to reflect the shortest transmission path.

[0140] When calculating the time difference between the delayed payment event and the supplier's supply disruption event in each related event pair, the unit of measurement is days. The date of the supplier's supply disruption event is subtracted from the date of the delayed payment event to obtain a set of transmission time difference values, which constitute a transmission time difference sequence.

[0141] When performing outlier removal on conduction time difference sequences, a method based on the interquartile range (interquartile range) can be used. The lower and upper quartiles of the conduction time difference sequence are calculated. Conduction time differences less than the lower quartile minus 1.5 times the interquartile range or greater than the upper quartile plus 1.5 times the interquartile range are identified as outliers and removed. Conduction time differences within a second pre-set confidence interval are retained, generating a valid conduction time difference sequence. The second pre-set confidence interval is the normal value range determined by the interquartile range method described above.

[0142] The arithmetic mean of the effective transmission time difference sequence is calculated. The sum of all values ​​in the sequence is then divided by the number of values ​​in the sequence to obtain the average transmission time difference. This value reflects the average time it takes for a delayed payment event to propagate to a supply disruption event. The reciprocal of the average transmission time difference is used as the first coupling strength value. The shorter the average transmission time difference and the larger the reciprocal, the more easily a disturbance in the funding subsystem triggers a rapid response in the materials subsystem, indicating a tighter coupling between the two subsystems.

[0143] This embodiment uses time window matching to filter pairs of suspected causally related events, such as delayed payment events and material supplier supply disruptions, from historical event data. It calculates the propagation time difference, removes outliers, takes the average, and then calculates the reciprocal to obtain the first coupling strength value. This achieves a quantitative assessment of the tightness of disturbance propagation between the funding subsystem and the material subsystem. This calculation process provides a key coupling strength input for the weighted summation of system resilience entropy.

[0144] In some embodiments, the vulnerability coefficients are weighted and summed with their corresponding coupling strength values ​​to generate the system resilience entropy, including:

[0145] Acquire historical shutdown event data, and extract the time-series change rate of each vulnerability coefficient and the time-series change rate of each coupling strength value before each historical shutdown event occurs from the historical shutdown event data;

[0146] The time-series rate of change of vulnerability coefficients and the time-series rate of change of coupling strength values ​​are normalized to generate normalized values ​​of each vulnerability rate of change and each coupling strength rate of change.

[0147] The joint sensitivity coefficient of each subsystem pair is generated by multiplying the normalized value of the change rate of each vulnerability with the normalized value of the change rate of the corresponding coupling strength.

[0148] The joint sensitivity coefficients are normalized to generate dynamic weights for each subsystem pair;

[0149] Each vulnerability coefficient is multiplied by its corresponding coupling strength value to generate the vulnerability-coupling product of each subsystem pair;

[0150] Multiply each fragile-coupling product by its corresponding dynamic weight to generate a weighted contribution value for each subsystem pair;

[0151] Sum all weighted contribution values ​​to generate the system resilience entropy.

[0152] In this embodiment, historical work stoppage event data is extracted from work stoppage reports or schedule delay records in the project management system. Each record contains information such as the occurrence time, duration, and attribution analysis of the work stoppage event. For each historical work stoppage event, time-series data of each vulnerability coefficient within a preset time period prior to the event are extracted. The difference in vulnerability coefficients between adjacent time points is calculated and divided by the time interval to obtain the time-series change rate of each vulnerability coefficient. Similarly, time-series data of each coupling strength value is extracted and its time-series change rate is calculated. These time-series change rates reflect the changing trends and fluctuations of each indicator before the work stoppage event. The time-series change rates of the vulnerability coefficients and coupling strength values ​​are normalized separately. The min-max normalization method is used to map both to the interval between 0 and 1, generating normalized values ​​for each vulnerability change rate and each coupling strength change rate to eliminate the impact of differences in the dimensions of different indicators on subsequent joint analysis.

[0153] The joint sensitivity coefficient of each subsystem pair is used to characterize the degree of coordinated change between the vulnerability and coupling strength of a subsystem pair before the occurrence of a historical shutdown event. A higher coefficient indicates a stronger correlation between the change in the subsystem pair and the shutdown event. The joint sensitivity coefficients are normalized so that the sum of the joint sensitivity coefficients of all subsystem pairs is 1, generating a dynamic weight for each subsystem pair. This weight reflects the relative importance of different subsystem pairs in historical shutdown events; a higher weight indicates a more significant impact of the change in the subsystem pair on the shutdown event. Because this weight is dynamically calculated based on historical shutdown event data, it can adapt to changes in the importance of subsystem pairs under different project stages and risk scenarios.

[0154] The vulnerability-coupling product of each subsystem pair reflects the combined level of the inherent vulnerability and the tightness of disturbance propagation of each subsystem pair at the current moment; the weighted contribution value of each subsystem pair reflects the actual contribution share of each subsystem pair in the system resilience entropy.

[0155] This embodiment introduces dynamic weights based on historical shutdown event data, enabling the calculation of system resilience entropy to adaptively adjust the importance of different subsystems in the aggregation process. Compared with the weighted summation method with fixed weights, it can more accurately reflect the actual contribution of each subsystem to the overall vulnerability of the system under different risk scenarios, thereby improving the effectiveness and sensitivity of system resilience entropy as an early warning indicator.

[0156] Please see Figure 3In some embodiments, the topology resilience coefficient is calculated based on the criticality of each process, the number of redundant paths, the total float time, and the coupling strength between high-impact processes in the construction schedule network diagram, including:

[0157] S301. Perform Monte Carlo simulation on each process in the construction plan network diagram, randomly disturb the duration of each process, and calculate the average impact of the duration disturbance of each process on the total duration. Use the average impact as the criticality of the process.

[0158] S302. Count the number of paths containing processes with a criticality lower than the preset criticality threshold in the construction plan network diagram, and use them as the number of redundant paths.

[0159] S303. Calculate the sum of the floating times of all processes in the construction schedule network diagram, and use it as the total floating time.

[0160] S304. For processes with criticality higher than the preset criticality threshold, pair them up in pairs, calculate the immediate predecessor and immediate successor time interval in the logical dependency relationship between each paired process, and use the reciprocal of the immediate predecessor and immediate successor time interval as the coupling strength between the paired processes.

[0161] S305. Select the maximum value among all coupling strengths as the coupling strength between high-impact processes;

[0162] S306. Add the number of redundant paths to the total floating time to generate the elastic gain term;

[0163] S307. Multiply the mean of criticality by the coupling strength between high-impact processes to generate an elastic loss term;

[0164] S308. The ratio of the elastic gain term to the elastic loss term is used as the topological elasticity coefficient.

[0165] In step S301, the Monte Carlo simulation assesses the impact of each process on the total project duration by applying random perturbations to its duration in the construction schedule network diagram. Specifically, a set of random duration perturbation values ​​following a specific probability distribution is generated for each process. The parameters of this probability distribution can be calibrated based on the historical duration deviation data of the process, for example, using the mean and standard deviation of historical duration deviations as parameters of a normal distribution. Each random perturbation value is superimposed on the planned duration of the process to obtain a set of perturbed duration values. These values ​​are then substituted into the construction schedule network diagram and the total project duration is recalculated along the logical dependencies. The deviation between the perturbed total project duration and the original planned total project duration is calculated, and the mean of all deviations is taken. This mean is the criticality of the process. A higher criticality indicates a greater impact of the process's duration fluctuations on the total project duration.

[0166] In step S302, a preset criticality threshold is used to distinguish between high-impact and low-impact processes. This threshold can be determined based on the distribution characteristics of the criticality of all processes, for example, by taking the median or upper quartile of all criticalities as the dividing point. The number of redundant paths is obtained by traversing and statistically analyzing all paths in the construction plan network diagram. For each path, if the criticality of all processes on that path is lower than the preset criticality threshold, then that path is counted as a redundant path. A higher number of redundant paths indicates more alternative construction routes in the construction plan, and better flexibility in the schedule network.

[0167] In step S303, the total float time is obtained by extracting the float time of each operation from the construction schedule network diagram and summing them up. Float time is the amount of time an operation can be delayed without affecting the overall project duration, and it is derived from the critical path analysis results of the construction schedule network diagram. The larger the total float time, the more time buffer exists in the schedule network, which can absorb some delays in operations without affecting the overall project duration.

[0168] In step S304, when pairing processes with a criticality higher than a preset criticality threshold, only process pairs with direct or indirect logical dependencies are considered, while process pairs without dependencies are ignored. The predecessor-follower time interval is extracted from the logical dependencies in the construction plan network diagram. For two processes with a completion-start dependency, the predecessor-follower time interval is the time difference between the completion time of the predecessor process and the start time of the follower process. The reciprocal of the predecessor-follower time interval is used as the coupling strength between the paired processes. The shorter the time interval, the higher the coupling strength, indicating that the two high-impact processes are more closely linked in time, and the more easily a delay in one process is propagated to the other.

[0169] In step S305, the maximum value among all the coupling strengths of the paired processes is selected as the coupling strength between the high-impact processes to reflect the tightest coupling relationship between the high-impact processes, i.e. the weakest transmission path.

[0170] In steps S306 to S308, the elasticity gain term quantifies the total amount of elastic factors in the schedule network that can absorb delays. The larger the elasticity gain term, the more alternative routes and time buffers the schedule network has, and the better it can absorb process delays. The elasticity loss term quantifies the total amount of vulnerable factors in the schedule network that cause delays to amplify. The larger the elasticity loss term, the greater the impact of high-impact processes on the overall project duration and the tighter their coupling, making it easier for local delays to trigger chain reactions. The ratio of the elasticity gain term to the elasticity loss term serves as the topological elasticity coefficient. The larger the ratio, the stronger the elasticity of the schedule network; the smaller the ratio, the more vulnerable the schedule network. When the topological elasticity coefficient is lower than a second preset threshold, it indicates that the schedule network has entered a vulnerable state, and any additional delay may cause the overall project duration to spiral out of control.

[0171] This embodiment quantifies the impact of each process on the total project duration through Monte Carlo simulation. By combining the number of redundant paths, the total floating time, and the coupling strength between high-impact processes, the ratio of elastic gain to elastic loss is constructed as the topological elasticity coefficient. This quantitatively evaluates the risk resistance of the construction plan from the perspective of the structural characteristics of the schedule network, providing another quantitative dimension that complements the system resilience entropy for subsequent early warning triggering.

[0172] In some embodiments, Monte Carlo simulations are performed on each process in the construction schedule network diagram, the duration of each process is randomly perturbed, and the average impact of the duration perturbed on the total duration is statistically analyzed. This average impact is used as the criticality of the process, including:

[0173] For each process in the construction plan network diagram, a sequence of random schedule disturbance values ​​that follows a preset probability distribution is generated. The preset probability distribution is calibrated based on the historical schedule deviation data of the process.

[0174] Each disturbance value in the random duration disturbance value sequence is sequentially superimposed onto the planned duration of the corresponding process to generate a set of disturbance-post duration values;

[0175] Substitute the disturbed project duration value into the construction schedule network diagram, recalculate the critical path along the logical dependencies in the construction schedule network diagram, and generate the disturbed total project duration corresponding to each disturbance value.

[0176] Calculate the difference between the total project duration after the disturbance and the original planned total project duration, and generate the project duration deviation corresponding to each disturbance value;

[0177] The average value of all schedule deviations is calculated to generate the average impact of each process on the total schedule, which is used as the criticality of the process.

[0178] In this embodiment, the type and parameters of the preset probability distribution are calibrated based on historical schedule deviation data of the process. Historical schedule deviation data is extracted from the historical execution records of the process in the project management system. The difference between the actual and planned durations of each historical execution batch is statistically analyzed to form a schedule deviation sample set. The probability distribution type is selected based on the distribution characteristics of the sample set. If the schedule deviation data exhibits a symmetrical distribution, a normal distribution is fitted; if it exhibits a skewed distribution, a triangular or log-normal distribution is fitted. The distribution parameters are calibrated using maximum likelihood estimation or moment estimation methods. Based on the calibrated preset probability distribution, a sequence of random schedule disturbance values ​​is generated through random sampling. The sequence length is set according to the required simulation accuracy, and each value in the sequence represents the magnitude of the random disturbance to the schedule of the process in a single simulation.

[0179] Each disturbance value in the random duration disturbance value sequence is sequentially superimposed onto the planned duration of the corresponding process, resulting in a set of disturbance-post-duration duration values. Each disturbance-post-duration duration value corresponds to an independent simulation scenario. These disturbance-post-duration duration values ​​are then substituted into the construction schedule network diagram, replacing the original planned duration of the corresponding process. The critical path method is then used to recalculate the duration of all paths along the logical dependencies in the construction schedule network diagram, identifying the new critical path and determining the total disturbance-post-duration duration. This process is repeated until all disturbance values ​​have been calculated, resulting in a set of total disturbance-post-duration duration values.

[0180] The difference between the total project duration after each disturbance and the original planned total project duration is calculated to obtain the schedule deviation. This value reflects the actual impact of the schedule disturbance of that process on the total project duration in this simulation scenario. The arithmetic mean of all schedule deviations is calculated to obtain the average impact of that process on the total project duration, which is taken as the criticality of the process. This criticality reflects the statistically significant average impact of the schedule fluctuation of that process on the total project duration; the higher the criticality, the greater the risk of the schedule uncertainty of that process to the overall project schedule.

[0181] This embodiment calibrates the probability distribution based on historical data and generates random disturbance values. Monte Carlo simulations are then performed on each process step, and the average impact of the duration disturbance on the total duration is used as the criticality of the process. This achieves a quantitative assessment of the importance of each process in the construction schedule network diagram. This criticality provides a core input parameter for the subsequent calculation of the topological resilience coefficient, enabling the topological resilience coefficient to accurately reflect the structural vulnerability of the construction schedule network.

[0182] In some embodiments, the number of redundant paths is added to the total float time to generate a resilience gain term; the mean criticality is multiplied by the coupling strength between high-impact processes to generate a resilience loss term; the ratio of the resilience gain term to the resilience loss term is used as the topology resilience coefficient, including:

[0183] Calculate the criticality of all processes in the construction plan network diagram, calculate the arithmetic mean of all criticalities, and generate the average criticality.

[0184] Multiply the average criticality by the coupling strength between high-impact processes to generate an elastic loss term;

[0185] Calculate the floating time of all processes in the construction plan network diagram, sum the floating time of each process, and generate the total floating time.

[0186] The number of paths containing processes with a criticality lower than a preset criticality threshold in the construction plan network diagram is counted as the number of redundant paths.

[0187] Add the number of redundant paths to the total floating time to generate the elastic gain term;

[0188] Calculate the ratio of the elastic gain term to the elastic loss term, and use the ratio as the topological elasticity coefficient.

[0189] In this embodiment, the average criticality is obtained by calculating the arithmetic mean of the criticalities of all processes in the construction schedule network diagram. This reflects the central tendency of the impact of each process on the overall project duration. A higher average criticality indicates that there are more processes in the construction schedule network diagram that have a significant impact on the overall project duration, and the schedule network is more vulnerable. Multiplying the average criticality by the coupling strength between high-impact processes generates an elasticity loss term. This product combines the average level of process importance with the tightness of coupling between high-impact processes. A larger elasticity loss term indicates a stronger vulnerability factor in the schedule network that leads to amplified delays.

[0190] The total float time is obtained by summing the float times of all processes in the construction schedule network diagram. The float time of each process is derived from the critical path analysis results of the construction schedule network diagram, reflecting the amount of time that a process can be delayed without affecting the overall project duration. The number of redundant paths is obtained by traversing all paths in the construction schedule network diagram and counting the number of paths containing processes with a criticality lower than a preset criticality threshold, reflecting the richness of alternative construction routes in the construction plan. The number of redundant paths is added to the total float time to generate a resilience gain term. This sum combines the number of alternative routes and the total time buffer. The larger the resilience gain term, the stronger the resilience of the schedule network in absorbing delays.

[0191] Calculate the ratio of the elastic gain term to the elastic loss term, and use this ratio as the topological elasticity coefficient. When the elastic gain term is relatively large and the elastic loss term is relatively small, the topological elasticity coefficient is high, indicating that the schedule network has sufficient elasticity to absorb process delays and the network structure is relatively robust. When the elastic gain term is relatively small and the elastic loss term is relatively large, the topological elasticity coefficient is low, indicating that the schedule network lacks elasticity and local delays can easily trigger a chain reaction leading to loss of control over the overall project schedule.

[0192] This embodiment constructs a topological elasticity coefficient by calculating the average criticality, total floating time, and number of redundant paths, and establishes the ratio of elastic gain to elastic loss. It clarifies the calculation method and physical meaning of each intermediate parameter, providing a specific implementation path for quantitatively evaluating the risk resistance of a construction plan from the perspective of schedule network structure characteristics.

[0193] Please see Figure 4 In some embodiments, when the system resilience entropy exceeds a first preset threshold or the topological elasticity coefficient is lower than a second preset threshold, a warning signal is generated, including:

[0194] S401. Obtain historical shutdown event data, extract the system resilience entropy value corresponding to the time of occurrence of each historical shutdown event from the historical shutdown event data, and generate a historical resilience entropy sequence.

[0195] S402. Extract the topological elasticity coefficient value corresponding to the time of occurrence of each historical shutdown event from the historical shutdown event data, and generate a historical elasticity coefficient sequence.

[0196] S403. Perform percentile statistics on the historical resilience entropy sequence, and use the resilience entropy value corresponding to the Nth percentile as the first preset threshold.

[0197] S404. Perform percentile statistics on the historical elasticity coefficient sequence, and use the elasticity coefficient value corresponding to the Mth percentile as the second preset threshold.

[0198] S405. Monitor the system resilience entropy at the current moment in real time and determine whether the system resilience entropy is greater than the first preset threshold.

[0199] If the system resilience entropy is greater than the first preset threshold, a first warning signal is generated, which indicates the risk of shutdown due to insufficient system resilience.

[0200] S406. Monitor the topology elasticity coefficient at the current moment in real time and determine whether the topology elasticity coefficient is less than the second preset threshold.

[0201] If the topology resilience coefficient is less than the second preset threshold, a second early warning signal is generated, which indicates the risk of work stoppage caused by the fragility of the progress network.

[0202] If the system resilience entropy is greater than the first preset threshold and the topology elasticity coefficient is less than the second preset threshold, a third warning signal is generated. The third warning signal indicates the risk of work stoppage caused by the superposition of insufficient system resilience and vulnerability of the schedule network.

[0203] In step S401, historical work stoppage event data is extracted from the work stoppage records of the project management system. During extraction, work stoppage events need to be screened and identified, excluding those caused by force majeure or unplanned factors such as weather, holiday arrangements, or planned material rotation. Only unplanned work stoppage events caused by systemic reasons such as resource shortages, management errors, or lack of coordination are retained to ensure that the data used for subsequent threshold calibration accurately reflects the risks arising from the inherent vulnerabilities of the construction system. For each identified historical work stoppage event, the system resilience entropy value corresponding to the time of the event is extracted from the historical calculation log of system resilience entropy, arranged chronologically to generate a historical resilience entropy sequence.

[0204] In step S402, similarly, the topology elasticity coefficient value corresponding to the time of occurrence of each identified historical shutdown event is extracted from the historical calculation log of the topology elasticity coefficient, and arranged in chronological order to generate a historical elasticity coefficient sequence.

[0205] In step S403, when performing percentile statistics on the historical resilience entropy sequence, all values ​​in the sequence are arranged in ascending order, and the value corresponding to the Nth percentile position is calculated as the first preset threshold. Since a higher system resilience entropy indicates a more fragile system, the warning trigger condition is that the system resilience entropy exceeds the first preset threshold. The first preset threshold should be taken from the region with higher values ​​in the historical resilience entropy sequence, that is, N in the Nth percentile should be a larger value, such as the 90th percentile or the 95th percentile, so that only a few historical shutdown events occur at times when the system resilience entropy value is higher than this threshold, ensuring the targeting of the warning. The specific value of N can be determined by traversing the candidate percentiles, calculating the corresponding false alarm rate and false negative rate respectively, and selecting the percentile with the smallest sum of the two.

[0206] In step S404, percentile statistics are performed on the historical resilience coefficient sequence. The resilience coefficient value corresponding to the Mth percentile is used as the second preset threshold. Since a smaller topological resilience coefficient indicates a more fragile progress network, the warning trigger condition is that the topological resilience coefficient is lower than the second preset threshold. The second preset threshold should be taken from the region with lower values ​​in the historical resilience coefficient sequence, that is, M in the Mth percentile should be a small value, such as the 10th percentile or the 5th percentile, so that only a few historical shutdown events occur at times when the topological resilience coefficient value is lower than this threshold, ensuring the targeting of the warning. The specific value of M is also determined by traversing the candidate percentiles, calculating the corresponding false alarm rate and false negative rate respectively, and selecting the percentile with the smallest sum of the two.

[0207] In steps S405 and S406, if the system resilience entropy is greater than a first preset threshold, it indicates that the overall resistance to disturbances of the current construction system has decreased to a dangerous level similar to that of historical shutdown events. At this time, a first warning signal is generated. This signal includes the current value of the system resilience entropy, the magnitude of the exceedance of the threshold, and the identifier of the subsystem with the highest contribution, used to alert management personnel to the risk of shutdown due to insufficient system resilience. If the topological elasticity coefficient is less than a second preset threshold, it indicates that the structural robustness of the current construction progress network has decreased to a vulnerable level similar to that of historical shutdown events. At this time, a second warning signal is generated. This signal includes the current value of the topological elasticity coefficient, the magnitude of the fall below the threshold, and the identifier of the most vulnerable critical process, used to alert management personnel to the risk of shutdown due to the vulnerability of the progress network. If the system resilience entropy is greater than the first preset threshold and the topological elasticity coefficient is less than the second preset threshold, a third warning signal is generated. This signal indicates that insufficient system resilience and vulnerability of the progress network are superimposed, significantly increasing the risk of shutdown, requiring priority handling.

[0208] This embodiment filters and identifies historical shutdown events to eliminate interference from non-systematic factors, making the data used for threshold calibration purer. By selecting percentiles from the high-order region of the historical resilience entropy sequence and the low-order region of the elasticity coefficient sequence as thresholds, it adapts to the different warning triggering directions of the two indicators. This embodiment combines parallel monitoring of dual thresholds with the differentiated generation of three warning signals, providing managers with precise alerts for different risk scenarios, improving the accuracy and operability of warnings.

[0209] In some embodiments, percentile statistics are performed on the historical resilience entropy sequence, and the resilience entropy value corresponding to the Nth percentile is used as a first preset threshold, including:

[0210] Obtain all resilience entropy values ​​from the historical resilience entropy sequence, sort all resilience entropy values ​​in ascending order, and generate an ordered resilience entropy sequence.

[0211] Calculate the length of the ordered resilience entropy sequence, multiply the length by the preset percentile value, and generate the index position value;

[0212] Round the index value down to generate an integer index;

[0213] Extract the resilience entropy value at the integer index position from the ordered resilience entropy sequence as a candidate threshold;

[0214] Calculate the false alarm rate and false negative rate corresponding to the candidate threshold. The false alarm rate is the probability that the system resilience entropy exceeds the candidate threshold when no shutdown event occurs, and the false negative rate is the probability that the system resilience entropy does not exceed the candidate threshold when a shutdown event occurs.

[0215] The candidate threshold corresponding to the minimum sum of false positive rate and false negative rate is taken as the first preset threshold.

[0216] In this embodiment, the selection range of the preset percentile values ​​covers multiple candidate values ​​from the low digit to the high digit. For example, the values ​​between 0.50 and 0.99 are traversed with a step size of 0.05, and each preset percentile value corresponds to a candidate threshold.

[0217] The calculation of the false alarm rate relies on the monitoring data of system resilience entropy during the period when no shutdown event occurred. This data is extracted from the historical calculation log of system resilience entropy, and the system resilience entropy values ​​corresponding to the time points that are not related to the time of the historical shutdown event are selected to form a negative sample set. The calculation of the false alarm rate directly uses the system resilience entropy values ​​corresponding to the time of the historical shutdown event to form a positive sample set.

[0218] For each candidate threshold, the proportion of negative samples exceeding the threshold is taken as the false alarm rate, and the proportion of positive samples not exceeding the threshold is taken as the false negative rate. The criterion of minimizing the sum of the false alarm rate and the false negative rate essentially seeks a balance between the two types of errors. When the sum of the two is minimized, the total error rate of the early warning decision corresponding to the candidate threshold is minimized. The candidate threshold that meets this condition is taken as the first preset threshold, so that the early warning trigger will not cause frequent false alarms due to the threshold being too low, nor will it miss real risks due to the threshold being too high.

[0219] This embodiment achieves automated calibration of the first preset threshold by traversing candidate percentiles and selecting the optimal threshold based on the criterion of minimizing the sum of false positive rate and false negative rate, thus avoiding the subjective bias of manually setting the threshold based on experience.

[0220] Please see Figure 5 In a second aspect, this embodiment also provides a key project construction progress and shutdown risk early warning system 1, applicable to the method described in the first aspect. The system includes a data acquisition module 11, a resilience entropy calculation module 12, an elasticity coefficient calculation module 13, an early warning generation module 14, a counterfactual reasoning module 15, and an output module 16. The data acquisition module 11 is used to acquire buffer capacity data of the labor subsystem, machinery subsystem, material subsystem, funding subsystem, and management subsystem of the construction system, as well as the logical dependencies and planned durations of each process in the construction plan network diagram. The resilience entropy calculation module 12 is connected to the data acquisition module 11 and is used to calculate the system resilience entropy based on the buffer capacity data of each subsystem and the coupling strength data between each subsystem. The elasticity coefficient calculation module 13 calculates the system resilience entropy. The calculation module 13 is connected to the data acquisition module 11 and is used to calculate the topological elasticity coefficient based on the criticality of each process, the number of redundant paths, the total floating time, and the coupling strength between high-impact processes in the construction plan network diagram. The early warning generation module 14 is connected to the resilience entropy calculation module 12 and the elasticity coefficient calculation module 13 respectively and is used to generate an early warning signal when the system resilience entropy exceeds the first preset threshold or the topological elasticity coefficient is lower than the second preset threshold. The counterfactual reasoning module 15 is connected to the early warning generation module 14 and is used to generate at least one intervention plan based on the structural causal model of the subsystem buffer capacity data in response to the early warning signal. The intervention plan includes the target subsystem, the target buffer capacity adjustment amount, and the expected effect. The output module 16 is connected to the counterfactual reasoning module 15 and is used to output the intervention plan and the expected effect.

[0221] In this system, the data acquisition module 11 collects buffer capacity data from the labor, equipment, materials, finance, and approval modules of the construction management information system, and obtains planned network data from the project management tool; the resilience entropy calculation module 12 and the elasticity coefficient calculation module 13 process the above data to output the system resilience entropy and topological elasticity coefficient; the early warning generation module 14 compares the two indicators with their respective thresholds and generates an early warning signal when the triggering conditions are met; the counterfactual reasoning module 15 performs intervention analysis on the buffer capacity data based on the structural causal model, generating an intervention plan that includes the target subsystem, adjustment amount, and expected effect; the output module 16 presents the intervention plan to the management interface. These modules are sequentially connected to form a data processing link from data acquisition to decision support, realizing automated closed-loop management of construction progress and work stoppage risk early warning.

[0222] By adopting the above technical solutions, this invention differs from existing technologies and has the following beneficial effects: The above technical solutions provide a method and system for early warning of construction progress and work stoppage risks in key projects. By acquiring buffer capacity data of the labor subsystem, machinery subsystem, material subsystem, capital subsystem, and management subsystem in the construction system, and the logical dependencies and planned durations of each process in the construction plan network diagram, the system resilience entropy is calculated based on the buffer capacity data of each subsystem and the coupling strength data between each subsystem. The topological elasticity coefficient is calculated based on the criticality of the process, the number of redundant paths, the total floating time, and the coupling strength between high-impact processes. When the system resilience entropy exceeds a first preset threshold or the topological elasticity coefficient is lower than a second preset threshold, an early warning signal is generated. In response to the early warning signal, counterfactual reasoning is performed on the subsystem buffer capacity data based on a structural causal model to generate and output an intervention plan that includes the target subsystem, the target buffer capacity adjustment amount, and the expected effect, thus realizing the automatic generation of advanced early warning and intervention plans. By collecting buffer capacity data from five subsystems—labor, machinery, materials, capital, and management—a system resilience entropy is constructed to quantify the overall vulnerability of the construction system to disturbances. Simultaneously, a topological elasticity coefficient is constructed based on the criticality of each process, the number of redundant paths, the total floating time, and the coupling strength between high-impact processes in the construction plan network diagram to quantify the structural robustness of the schedule network. These two indicators provide an advanced assessment of downtime risk from both system constitution and network structure dimensions. A dual-threshold parallel monitoring mechanism combining system resilience entropy and topological elasticity coefficients is employed. When either indicator exceeds its limit, a differentiated early warning signal is triggered, transforming the risk status into an identifiable management signal. In response to the early warning signal, counterfactual reasoning is performed on the subsystem buffer capacity data based on a structural causal model, automatically generating and outputting an intervention plan that includes the target subsystem, the target buffer capacity adjustment amount, and the expected effect. This achieves a complete closed loop from risk identification to decision support. This invention transforms passive, reactive progress monitoring into proactive, preventative constitution assessment and further provides quantifiable intervention suggestions, significantly improving the intelligence level and decision-making efficiency of key project construction progress and downtime risk management.

[0223] Finally, it should be noted that although the above embodiments have been described in the text and drawings of this application, this should not limit the scope of patent protection of this application. Any technical solutions that are based on the essential concept of this application and utilize the content described in the text and drawings of this application, resulting in equivalent structural or procedural substitutions or modifications, as well as the direct or indirect application of the technical solutions of the above embodiments to other related technical fields, are all included within the scope of patent protection of this application.

Claims

1. A method for early warning of construction progress and work stoppage risks in key projects, characterized in that, include: Obtain buffer capacity data for the labor subsystem, machinery subsystem, materials subsystem, funding subsystem, and management subsystem of the construction system; Obtain the logical dependencies and planned durations of each process in the construction schedule network diagram; Based on the buffer capacity data of each subsystem and the coupling strength data between each subsystem, the system resilience entropy is calculated, including: The buffer capacity data of the labor subsystem, the machinery subsystem, the materials subsystem, the funds subsystem, and the management subsystem are normalized respectively to generate normalized buffer capacity values ​​for each subsystem. Based on the transmission time interval between fund delay payment events and material supplier supply disruption events in historical event data, the first coupling strength value between the fund subsystem and the material subsystem is determined. Based on the transmission time interval between fund delay payment events and subcontractor passive work events in the historical event data, a second coupling strength value between the fund subsystem and the labor subsystem is determined. Based on the transmission time interval between material shortage events and mechanical downtime events in the historical event data, a third coupling strength value between the material subsystem and the mechanical subsystem is determined. Based on the transmission time interval between worker absence events and machine operator shutdown events in the historical event data, a fourth coupling strength value between the labor subsystem and the mechanical subsystem is determined. Based on the set of transmission time intervals between approval delay events and response delay events of each subsystem in the historical event data, a fifth coupling strength value between the management subsystem and each other subsystem is determined. The vulnerability coefficient of each subsystem is generated by taking the reciprocal of each normalized buffer capacity value. The system resilience entropy is generated by weighted summation of each vulnerability coefficient and its corresponding coupling strength value. Based on the criticality of each process, the number of redundant paths, the total floating time, and the coupling strength between high-impact processes in the construction plan network diagram, the topology elasticity coefficient is calculated. When the system resilience entropy exceeds a first preset threshold or the topological elasticity coefficient is lower than a second preset threshold, an early warning signal is generated. In response to the warning signal, counterfactual reasoning is performed on the subsystem buffer capacity data based on a structural causal model to generate at least one intervention plan, which includes the target subsystem, the target buffer capacity adjustment amount, and the expected effect. Output the intervention plan and the expected results.

2. The method for early warning of construction progress and work stoppage risk in key projects according to claim 1, characterized in that, Based on the transmission time interval between delayed payment events and material supplier supply disruption events in historical event data, the first coupling strength value between the funding subsystem and the material subsystem is determined, including: Extract the timestamp sequences of fund delay payment events and material supplier supply disruption events from the historical event data; The timestamp sequence of the fund delay payment event is matched with the timestamp sequence of the material supplier supply interruption event to determine the material supplier supply interruption event that occurs within the first preset time window after each fund delay payment event, as a related event pair. Calculate the time difference between the delayed payment event and the material supplier supply disruption event in each of the aforementioned related event pairs to generate a transmission time difference sequence; The conduction time difference sequence is subjected to outlier removal processing, and the conduction time difference within the second preset information interval is retained to generate a valid conduction time difference sequence; Calculate the mean of the effective conduction time difference sequence to generate the average conduction time difference; The reciprocal of the average conduction time difference is used as the first coupling strength value.

3. The method for early warning of construction progress and work stoppage risk in key projects according to claim 1, characterized in that, The system resilience entropy is generated by weighted summation of each vulnerability coefficient and its corresponding coupling strength value, including: Acquire historical shutdown event data, and extract the time-series change rate of each vulnerability coefficient and the time-series change rate of each coupling strength value before each historical shutdown event occurs from the historical shutdown event data; The time-series change rate of the vulnerability coefficient and the time-series change rate of the coupling strength value are normalized respectively to generate normalized values ​​of each vulnerability change rate and each coupling strength change rate. The joint sensitivity coefficient of each subsystem pair is generated by multiplying the normalized value of the change rate of each vulnerability by the normalized value of the change rate of the corresponding coupling strength. The joint sensitivity coefficients are normalized to generate dynamic weights for each subsystem pair; Each vulnerability coefficient is multiplied by its corresponding coupling strength value to generate a vulnerability-coupling product for each subsystem pair; Multiply each of the fragile-coupling products by the corresponding dynamic weights to generate a weighted contribution value for each subsystem pair; The system resilience entropy is generated by summing all the weighted contribution values.

4. The method for early warning of construction progress and work stoppage risk in key projects according to claim 1, characterized in that, Based on the criticality of each process, the number of redundant paths, the total float time, and the coupling strength between high-impact processes in the construction plan network diagram, the topology flexibility coefficients are calculated, including: Monte Carlo simulation is performed on each process in the construction plan network diagram. The duration of each process is randomly disturbed, and the average impact of the duration disturbance of each process on the total duration is calculated. The average impact is used as the criticality of the process. The number of paths containing processes with a criticality lower than a preset criticality threshold in the construction plan network diagram is counted and used as the number of redundant paths; The sum of the floating times of all processes in the construction plan network diagram is taken as the total floating time. For processes with a criticality higher than the preset criticality threshold, pair them up in pairs, calculate the immediate preceding and immediate following time intervals in the logical dependency relationship between each paired process, and use the reciprocal of the immediate preceding and immediate following time intervals as the coupling strength between the paired processes. The maximum value among all the coupling strengths is selected as the coupling strength between the high-impact processes; Add the number of redundant paths to the total floating time to generate an elastic gain term; The mean of the criticality is multiplied by the coupling strength between the high-impact processes to generate an elastic loss term; The ratio of the elastic gain term to the elastic loss term is used as the topological elasticity coefficient.

5. The method for early warning of construction progress and work stoppage risk in key projects according to claim 4, characterized in that, Monte Carlo simulations are performed on each process in the construction schedule network diagram, randomly perturbing the duration of each process. The average impact of the duration perturbation on the total duration is calculated, and this average impact is used as the criticality of the process, including: For each process in the construction plan network diagram, a sequence of random schedule disturbance values ​​that follows a preset probability distribution is generated, the preset probability distribution being calibrated based on the historical schedule deviation data of the process; Each disturbance value in the random duration disturbance value sequence is sequentially superimposed onto the planned duration of the corresponding process to generate a set of disturbance-post duration values; Substitute the disturbed project duration value into the construction plan network diagram, and recalculate the critical path along the logical dependencies in the construction plan network diagram to generate the disturbed total project duration corresponding to each disturbance value. Calculate the difference between the total project duration after the disturbance and the original planned total project duration, and generate the project duration deviation corresponding to each disturbance value; The average value of all the aforementioned schedule deviations is calculated to generate the average impact of the process on the total schedule, which is used as the criticality of the process.

6. The method for early warning of construction progress and work stoppage risk in key projects according to claim 4, characterized in that, The number of redundant paths is added to the total floating time to generate an elastic gain term; the mean of the criticality is multiplied by the coupling strength between the high-impact processes to generate an elastic loss term. The ratio of the elastic gain term to the elastic loss term is used as the topological elasticity coefficient, including: The criticality of all processes in the construction plan network diagram is statistically analyzed, and the arithmetic mean of all criticalities is calculated to generate the average criticality. The elastic loss term is generated by multiplying the average criticality by the coupling strength between the high-impact processes. The floating time of all processes in the construction plan network diagram is calculated, and the floating time of each process is summed to generate the total floating time. The number of paths containing processes with a criticality lower than a preset criticality threshold in the construction plan network diagram is counted and used as the number of redundant paths; The number of redundant paths is added to the total floating time to generate the elastic gain term; Calculate the ratio of the elastic gain term to the elastic loss term, and use the ratio as the topological elasticity coefficient.

7. The method for early warning of construction progress and work stoppage risk in key projects according to claim 1, characterized in that, When the system resilience entropy exceeds a first preset threshold or the topological elasticity coefficient falls below a second preset threshold, a warning signal is generated, including: Acquire historical shutdown event data, extract the system resilience entropy value corresponding to the time of occurrence of each historical shutdown event from the historical shutdown event data, and generate a historical resilience entropy sequence; Extract the topological elasticity coefficient value corresponding to the time of occurrence of each historical shutdown event from the historical shutdown event data to generate a historical elasticity coefficient sequence; Perform percentile statistics on the historical resilience entropy sequence, and use the resilience entropy value corresponding to the Nth percentile as the first preset threshold. Perform percentile statistics on the historical elasticity coefficient sequence, and use the elasticity coefficient value corresponding to the Mth percentile as the second preset threshold; The system resilience entropy at the current moment is monitored in real time, and it is determined whether the system resilience entropy is greater than the first preset threshold. If the system resilience entropy is greater than the first preset threshold, a first warning signal is generated, which indicates the risk of shutdown due to insufficient system resilience. The topology elasticity coefficient is monitored in real time, and it is determined whether the topology elasticity coefficient is less than the second preset threshold. If the topology resilience coefficient is less than the second preset threshold, a second early warning signal is generated, which indicates the risk of work stoppage caused by the vulnerability of the progress network. If the system resilience entropy is greater than the first preset threshold and the topology elasticity coefficient is less than the second preset threshold, a third early warning signal is generated. The third early warning signal indicates the risk of work stoppage caused by the superposition of insufficient system resilience and vulnerability of the schedule network.

8. The method for early warning of construction progress and work stoppage risk in key projects according to claim 7, characterized in that, Perform percentile statistics on the historical resilience entropy sequence, and use the resilience entropy value corresponding to the Nth percentile as the first preset threshold, including: Obtain all resilience entropy values ​​in the historical resilience entropy sequence, and sort all the resilience entropy values ​​in ascending order to generate an ordered resilience entropy sequence. Calculate the length of the ordered resilience entropy sequence, multiply the length by a preset percentile value, and generate an index position value; The index position value is rounded down to generate an integer index; Extract the resilience entropy value at the integer index position from the ordered resilience entropy sequence as a candidate threshold; Calculate the false alarm rate and false negative rate corresponding to the candidate threshold, where the false alarm rate is the probability that the system resilience entropy exceeds the candidate threshold when no shutdown event occurs, and the false negative rate is the probability that the system resilience entropy does not exceed the candidate threshold when a shutdown event occurs. The candidate threshold corresponding to the minimum sum of the false positive rate and the false negative rate is taken as the first preset threshold.

9. A construction progress and work stoppage risk early warning system for key projects, characterized in that, The system applicable to the method of any one of claims 1 to 8, the system comprising: The data acquisition module is used to acquire buffer capacity data of the labor subsystem, machinery subsystem, material subsystem, financial subsystem, and management subsystem of the construction system, as well as the logical dependencies and planned duration of each process in the construction plan network diagram. A resilience entropy calculation module, connected to the data acquisition module, is used to calculate the system resilience entropy based on the buffer capacity data of each subsystem and the coupling strength data between each subsystem. The elasticity coefficient calculation module is connected to the data acquisition module and is used to calculate the topology elasticity coefficient based on the criticality of each process, the number of redundant paths, the total floating time, and the coupling strength between high-impact processes in the construction plan network diagram. The early warning generation module is connected to the resilience entropy calculation module and the elasticity coefficient calculation module respectively, and is used to generate an early warning signal when the system resilience entropy exceeds a first preset threshold or the topological elasticity coefficient is lower than a second preset threshold; The counterfactual reasoning module, connected to the early warning generation module, is used to respond to the early warning signal by performing counterfactual reasoning on the subsystem buffer capacity data based on a structural causal model to generate at least one intervention plan. The intervention plan includes the target subsystem, the target buffer capacity adjustment amount, and the expected effect. The output module, connected to the counterfactual reasoning module, is used to output the intervention plan and the expected effect.

Images