An intelligent project scheduling system and method based on engineering management

Abstract

The application belongs to the technical field of engineering management, and provides an intelligent project scheduling system and method based on engineering management. A plurality of candidate scheduling schemes are generated based on project space-time attribute graphs, task requirements and constraint conditions. For each scheme, the qualification time limit data of resources and the information flow correlation data between tasks are extracted, a risk deduction model is constructed and Monte Carlo simulation is performed, and the deduction results including risk probability distribution and risk root cause chain are output. The risk, cost and duration indexes of each scheme are input into a multi-objective optimization model for solution, and a Pareto optimal scheduling scheme is generated. Real-time data is dynamically collected during execution to update the risk model, and re-scheduling is triggered when the risk exceeds the threshold or the progress exceeds the limit. Finally, the scheduling process is restarted based on real-time state and risk analysis. The application improves the overall robustness and management adaptability of complex engineering projects in terms of cost, duration and compliance.

Description

Technical Field

[0001] This invention relates to the field of engineering management technology, and in particular to an intelligent project scheduling system and method based on engineering management. Background Technology

[0002] In complex engineering fields, project scheduling faces unprecedented challenges. These projects typically involve multiple heterogeneous resources (such as personnel with specific confidentiality qualifications and specialized equipment), complex logical and informational dependencies between tasks, and strict cost and schedule constraints. Traditional project scheduling methods, such as tools based on the Critical Path Method (CPM) or Project Review and Approval Technique (PERT), suffer from a core flaw in their static and deterministic assumptions. They treat task durations and resource availability as fixed values, failing to effectively handle two types of dynamic core risks: 1. Time-based qualification and continuity risks: The professional qualifications of project members (such as military confidentiality qualifications) have an expiration date and may become invalid during project execution. Traditional methods cannot proactively simulate the probability of such events occurring at a future point in time and the chain reactions they may cause (such as mission interruption, cost surges due to emergency personnel replacement, and schedule delays) when formulating the initial plan.

[0003] 2. Information security and compliance risks based on association: In classified or multi-classified projects, the improper flow of information along the task collaboration network is a significant hidden danger. Traditional scheduling tools only consider the time availability of resources and skill matching, lacking the ability to model and analyze the "task-person-information" association network, and cannot proactively embed and verify isolation mechanisms in the scheduling scheme to prevent the spread of information across classification levels.

[0004] Existing research on intelligent scheduling has introduced multi-objective optimization, but the optimization objectives are mostly limited to traditional dimensions such as schedule, cost, and resource balance. It fails to quantify the aforementioned dynamic compliance risks, continuity risks, and information security risks, and incorporate them as core optimization objectives and constraints into the decision-making model. This results in the generated "optimal" scheduling scheme being exceptionally fragile in actual implementation. Once encountering the aforementioned risks, it is highly susceptible to instability, forcing the initiation of costly passive responses and extensive rescheduling, severely undermining the scientific nature and controllability of project management.

[0005] Therefore, it is necessary to provide an intelligent project scheduling system and method based on engineering management to solve the above-mentioned technical problems. Summary of the Invention

[0006] To address the aforementioned technical issues, this invention provides an intelligent project scheduling system and method based on engineering management. Through the collaborative innovation of spatiotemporal attribute maps, risk extrapolation models, and multi-objective optimization models, it improves the scheduling efficiency and risk control capabilities of complex engineering projects in dynamic environments.

[0007] This invention provides an intelligent project scheduling method based on engineering management, the scheduling method comprising the following steps: Based on the pre-built spatiotemporal attribute map of the project, as well as the input project task requirements and constraints, multiple candidate scheduling schemes are generated. For each candidate scheduling scheme, a risk simulation model is constructed and a full-cycle risk simulation is performed, and the corresponding simulation results are output. The simulation results include the risk probability distribution, the risk root cause chain, the estimated cost, and the estimated construction period. The risk probability distribution, estimated cost, and estimated duration corresponding to each candidate scheduling scheme are input into a multi-objective optimization model for solution and trade-off, thereby generating the optimal scheduling scheme. During the execution of the optimal scheduling scheme, real-time status data is collected as new evidence and input into the risk inference model to update the risk probability distribution, and a rescheduling signal is triggered when the preset conditions are not met. In response to the rescheduling signal, a rescheduling process is triggered based on real-time status data, the updated risk probability distribution, and the associated risk root cause chain.

[0008] Preferably, the risks include at least one of compliance risks, continuity risks, and information security risks.

[0009] Preferably, the generation of multiple candidate scheduling schemes based on a pre-constructed project spatiotemporal attribute map and the input project task requirements and constraints includes: Obtain the resource status data from the project's spatiotemporal attribute map, the task attribute data from the project's task requirements, and the constraints. Based on the resource status data and the task attribute data, a multi-dimensional matching calculation is performed to generate an initial scheduling scheme set. The schemes in the initial scheduling scheme set are verified for compliance and feasibility based on the constraints, resulting in a verified scheme set. Based on a preset scheduling strategy, the schemes in the verified scheme set are adjusted and optimized to output the multiple candidate scheduling schemes.

[0010] Preferably, the preset scheduling strategy is at least one of the following: total cost minimization strategy, resource load balancing strategy, and task priority guarantee strategy.

[0011] Preferably, for each candidate scheduling scheme, a risk simulation model is constructed and a full-cycle risk simulation is performed, outputting the corresponding simulation results, including: For each candidate scheduling scheme, extract the corresponding resource qualification and timeliness data and the information flow association data between tasks from the project spatiotemporal attribute map; The risk simulation model is constructed based on the qualification timeliness data and the information flow association data. The risk simulation model is used to simulate the probability of qualification failure events occurring along the time axis and the impact of the task interruption it triggers, and to simulate the risk of information spreading across tasks along the path defined by the information flow association data. The risk simulation model is run multiple times for full-cycle simulation. In each simulation, random risk events are triggered based on the probability model, and the resulting task impact sequence, cumulative additional costs, and cumulative project delays are recorded. A causal aggregation analysis is performed on the task impact sequence recorded in multiple simulations to generate the risk root cause chain that reflects the risk transmission path; and the cumulative additional costs and cumulative project delays from multiple simulations are statistically analyzed to generate the risk probability distribution, the estimated cost, and the estimated project duration.

[0012] Preferably, the step of inputting the risk probability distribution, the estimated cost, and the estimated construction period corresponding to each candidate scheduling scheme into a multi-objective optimization model for solution and trade-off includes: For each candidate scheduling scheme, a comprehensive risk index value is extracted from the risk probability distribution and used together with the estimated cost and the estimated construction period as input parameters of the multi-objective optimization model. The weighting relationships between the risk objective function, cost objective function, and schedule objective function are set, and the comprehensive evaluation value of each candidate scheduling scheme is output based on the input parameters and the weighting relationships.

[0013] Preferably, generating the optimal scheduling scheme includes: All candidate scheduling schemes are sorted based on the comprehensive evaluation value, and schemes are selected according to a set ratio to form a Pareto front solution set. Conflict target analysis is performed on the schemes in the Pareto front solution set, and the final optimal scheduling scheme is selected from the Pareto front solution set according to the preset decision preferences.

[0014] Preferably, during the execution of the optimal scheduling scheme, collecting real-time status data as new evidence to input into the risk inference model, updating the risk probability distribution, and triggering a rescheduling signal when preset conditions are not met includes: During the execution of the optimal scheduling scheme, real-time status data reflecting task progress, resource status and external events are collected at a set sampling frequency. The real-time status data is used as time-series evidence and input into the risk inference model to drive the risk inference model to perform probability inference and output the updated risk probability distribution. The updated risk probability distribution is compared with the preset risk probability thresholds, and the progress deviation between the actual project progress and the estimated construction period is calculated. When the probability value of any risk in the updated risk probability distribution exceeds its corresponding risk probability threshold, or the progress deviation exceeds the preset tolerance range, it is determined that the preset conditions are not met, and the rescheduling signal is triggered.

[0015] Preferably, the rescheduling process, triggered in response to the rescheduling signal based on real-time status data, the updated risk probability distribution, and the associated risk root cause chain, includes: In response to the rescheduling signal, obtain a snapshot of the project site at the current moment; Using the project site snapshot as the new initial state, and combining the updated risk probability distribution and the risk root cause chain, adjust the project task requirements and constraints. Based on the adjusted project task requirements and constraints, as well as the pre-constructed project spatiotemporal attribute map, multiple candidate scheduling schemes are regenerated to trigger and complete the rescheduling process.

[0016] This invention also provides an intelligent project scheduling system based on engineering management, used to execute the aforementioned intelligent project scheduling method based on engineering management, the scheduling system comprising: The candidate solution generation module is used to generate multiple candidate scheduling solutions based on a pre-built spatiotemporal attribute map of the project and the input project task requirements and constraints. The cycle risk simulation module is used to construct a risk simulation model for each candidate scheduling scheme and perform a full cycle risk simulation, and output the corresponding simulation results, wherein the simulation results include risk probability distribution, risk root cause chain, estimated cost and estimated construction period; The multi-objective optimization decision module is used to input the risk probability distribution, the estimated cost and the estimated construction period corresponding to each candidate scheduling scheme into the multi-objective optimization model for solution and trade-off, and generate the optimal scheduling scheme. The rescheduling trigger module is used to collect real-time status data as new evidence to input into the risk inference model during the execution of the optimal scheduling scheme, update the risk probability distribution, and trigger a rescheduling signal when the preset conditions are not met. The rescheduling execution module is used to respond to the rescheduling signal and trigger the rescheduling process based on real-time status data, the updated risk probability distribution, and the related risk root cause chain.

[0017] Compared with related technologies, the intelligent project scheduling system and method based on engineering management provided by this invention has the following beneficial effects: This invention constructs a risk projection model that integrates qualification timeliness and information flow correlation data, and performs full-cycle Monte Carlo simulations on each candidate scheduling scheme. This allows for the quantitative assessment of the probability of occurrence, impact consequences, and transmission paths (risk root cause chains) of various risk events under different scheduling strategies during the planning phase. This transforms project management from a passive "encountering risk - emergency handling" model to an immune model of "anticipating risk - proactively avoiding it in scheme design," greatly enhancing the robustness and reliability of the plan.

[0018] This invention introduces "risk probability distribution" as a quantifiable objective that can be directly compared and weighed with cost and schedule, into a multi-objective optimization model. It can automatically generate a set of Pareto optimal solutions, clearly demonstrating the trade-offs between "lower risk," "lower cost," and "shorter schedule." This supports managers in making scientific decisions based on project phase preferences (such as prioritizing safety in the early stages and schedule in the later stages), thereby achieving true global comprehensive optimization, rather than local cost or time optimization.

[0019] This invention updates the risk simulation model by continuously injecting real-time status data during project execution, enabling the system to dynamically perceive changes in risk probability and schedule deviations. Once tolerance is exceeded, the system immediately triggers a rescheduling process and intelligently adjusts subsequent plans using the latest "risk probability distribution" and "risk root cause chain," achieving closed-loop adaptive control of "monitoring-early warning-diagnosis-re-optimization." This significantly improves the project's resilience and execution success rate in uncertain environments, while reducing management deviations and error correction costs. Attached Figure Description

[0020] Figure 1 A flowchart illustrating an intelligent project scheduling method based on engineering management provided by this invention; Figure 2 This invention provides a schematic diagram of the module structure of an intelligent project scheduling system based on engineering management. Detailed Implementation

[0021] The present invention will now be described in further detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and not intended to limit it. Furthermore, it should be noted that, for ease of description, only the parts relevant to the invention are shown in the drawings, not all structures. Moreover, unless otherwise specified, the embodiments and features described herein can be combined with each other.

[0022] It should also be noted that, for ease of description, the accompanying drawings show only the parts relevant to the invention and not all of them. Before discussing exemplary embodiments in more detail, it should be mentioned that some exemplary embodiments are described as processes or methods depicted as flowcharts. Although the flowcharts describe the operations (or steps) as sequential processes, many of the operations can be performed in parallel, concurrently, or simultaneously. Furthermore, the order of the operations can be rearranged. The process can be terminated when its operation is completed, but it may also have additional steps not included in the drawings. The process may correspond to a method, function, procedure, subroutine, subroutine, etc.

[0023] Example 1 This invention provides an intelligent project scheduling method based on engineering management, with reference to... Figure 1 As shown, the scheduling method includes the following steps: S1: Based on the pre-built spatiotemporal attribute map of the project, as well as the input project task requirements and constraints, generate multiple candidate scheduling schemes.

[0024] The risks include at least one of the following: compliance risks, continuity risks, and information security risks.

[0025] Specifically, step S1 includes the following steps: S11: Obtain the resource status data in the project spatiotemporal attribute map, the task attribute data in the project task requirements, and the constraints.

[0026] In this embodiment, firstly, real-time status data of all available resources at the current moment is extracted from the pre-constructed "project spatiotemporal attribute map". The resource status data is a structured dataset, which for each personnel resource includes at least their skill tag set (e.g., programming language, professional field), current workload status (whether idle or involved in other projects), affiliation (internal employee, outsourced personnel), geographical location, and key compliance attributes—especially the type, level, and specific effective and expiration dates of various professional qualification certificates. For non-human resources such as equipment and facilities, their type, performance indicators, current occupancy status, and available time window are extracted.

[0027] Secondly, the input "project task requirements" are parsed to extract task attribute data. This includes parsing the project's work breakdown structure (WBS) and generating an attribute set for each leaf task, which should include: the skill set required for the task, the estimated baseline working hours, the logical dependencies between tasks (complete-start, start-start, etc.), the task priority identifier, and, for specific tasks in classified projects, their information security level must be clearly defined.

[0028] Finally, define the project-level "constraints," which are usually input as global parameters. These mainly include: the upper limit of the total project cost budget, the final milestone date that the project must complete, the unit price of labor costs (which may vary depending on the type of personnel), and a library of mandatory policy rules that must be followed, such as "classified tasks must be performed by personnel holding the corresponding valid security clearance."

[0029] S12: Perform multi-dimensional matching calculations based on the resource status data and the task attribute data to generate an initial scheduling scheme set.

[0030] In this embodiment, the system performs automated multi-dimensional matching and searching within the "resource-task" two-dimensional space based on the data obtained in S11 to generate a batch of feasible preliminary solutions.

[0031] During implementation, a matching score calculation function can be designed to calculate a matching score for each "task-resource" pair. The score comprehensively considers the following core dimensions: skill fit (the degree of overlap between the skills required by the task and the skills possessed by the resource), time availability (whether the resource can be scheduled within the task's planned time period), and qualification compliance (whether the resource possesses the corresponding qualifications for tasks with confidentiality requirements).

[0032] The system employs heuristic search algorithms (such as priority-based list scheduling algorithms and genetic algorithms) to attempt to allocate the most suitable available resources from the resource pool for each task, prioritizing tasks based on their dependencies and priorities. It also assigns specific start and end times to each task, ensuring that resources do not conflict over time. By running the search algorithm multiple times, or by changing parameters such as task scheduling order and resource selection strategies, the system can generate a set of "initial scheduling schemes" that are feasible in basic dimensions such as skill, time, and qualifications, but may differ in cost and load distribution. Each initial scheme explicitly records who (resource) will execute each task and when (time window).

[0033] S13: Perform compliance and feasibility checks on the schemes in the initial scheduling scheme set according to the constraints, and obtain the checked scheme set.

[0034] In this embodiment, for each initial scheduling scheme generated in S12, the system performs a thorough verification based on the hard "constraints" explicitly defined in S11. Compliance verification is key, especially for confidentiality requirements. The system iterates through all tasks in the scheme, checking whether the personnel resources executing the task are within the task's execution time interval and whether their corresponding qualification certificates (such as confidentiality qualifications) are valid. If any task is found to be at risk of being executed "without qualification" or "with expired qualification," the scheme will be marked as non-compliant. Feasibility verification mainly checks for hard conflicts at the resource level, such as whether the same person is inappropriately assigned to handle two tasks at the same time, or whether the task arrangement violates the physical equipment's required maintenance interval. The system will check these constraints item by item. Any scheme that fails the verification will be removed from the set. Schemes that pass all verifications constitute the "verified scheme set." This step ensures that all subsequent optimized schemes are based on legality, compliance, and physical feasibility.

[0035] S14: Adjust and optimize the schemes in the verified scheme set based on the preset scheduling strategy, and output the multiple candidate scheduling schemes.

[0036] In this embodiment, the system further refines and optimizes the schemes in the "verified scheme set" with multi-objective guidance to generate high-quality "candidate scheduling schemes" with different advantages for subsequent risk simulation.

[0037] During implementation, the system loads one or more preset scheduling strategies. For example, if the "total cost minimization strategy" is enabled, the algorithm will attempt to replace high-cost resources with compliant resources at lower unit costs, or optimize task scheduling to reduce indirect costs caused by waiting time, without violating verification conditions. If the "resource load balancing strategy" is enabled, the algorithm will analyze the long-term load curves of each resource and, by fine-tuning task scheduling, appropriately distribute excessively dense tasks to smooth resource utilization. If the "task priority guarantee strategy" is enabled, the algorithm will ensure that high-priority tasks are not only allocated high-quality resources, but their scheduling will also be advanced as much as possible with buffers.

[0038] The optimization process typically employs algorithms such as local search and simulated annealing, iteratively improving each solution in the "verified solution set". Ultimately, the system outputs a set (usually dozens) of "candidate scheduling solutions" with distinct characteristics in different dimensions such as cost, load balancing, and mission-critical assurance, providing a wealth of input options for subsequent in-depth risk quantification analysis and comprehensive optimization.

[0039] The preset scheduling strategy is at least one of the following: total cost minimization strategy, resource load balancing strategy, and task priority guarantee strategy.

[0040] Specifically, the total cost minimization strategy aims to reduce the total project execution cost. During implementation, the system needs to define a clear cost model for each type of resource (especially human resources), such as the unit time cost of internal full-time employees, the unit time rate of outsourced personnel, and the rental or depreciation costs of equipment. During the optimization process, algorithms (such as local search or genetic algorithms) iteratively adjust the solutions in the "verified solution set." Typical operations include: attempting to replace higher-cost resources with compliant resources with lower unit costs, provided that skill and qualification requirements are met; optimizing task scheduling to reduce resource idle waiting time or shorten the total project duration, thereby reducing indirect costs; or adjusting resource input intensity (such as reducing parallelism to reduce management and coordination costs). The output of this strategy is a candidate solution that tends to minimize the sum of direct human and resource costs and indirect management costs.

[0041] Resource load balancing strategies aim to avoid resource overload or idleness, promoting smooth and stable resource utilization. During implementation, the system first calculates the load curve over time for each resource (such as key engineers or dedicated testing equipment) across various candidate solutions. The optimization algorithm aims to reduce the "peaks" (overload) and fill the "troughs" (idleness) of the load curve. Specific operations may include: appropriately staggering concurrent tasks densely allocated to a resource on the timeline; partially transferring overloaded tasks to other resources with similar skills and currently lighter loads; or fine-tuning the start time of non-critical tasks to smooth overall resource demand. This strategy typically establishes an objective function by minimizing the variance or maximum load of all resource loads. Its output is candidate solutions that significantly improve overall resource utilization and availability, reducing the risk of project delays due to bottlenecks in individual resources.

[0042] The task-priority strategy ensures that critical tasks essential to project success receive optimal resource allocation and scheduling. Before implementation, critical tasks (such as key technical breakthroughs and core integration testing) must be clearly identified in the project requirements. During optimization, this strategy assigns higher scheduling priority to critical tasks. Specific implementations include: prioritizing the allocation of resources with the highest skill matching, best performance, or highest reliability to critical tasks; scheduling critical tasks during periods of optimal resource availability and minimal disruption, and providing schedule buffers; and prioritizing the fulfillment of critical task needs when resource competition arises. The optimization focus of this strategy is not directly cost reduction or load balancing, but rather ensuring the high-quality and reliable execution of critical path or key milestone tasks to guarantee the achievement of overall project goals. Its output solution demonstrates superior success rate and timeliness in critical tasks.

[0043] S2: For each candidate scheduling scheme, construct a risk simulation model and perform a full-cycle risk simulation, and output the corresponding simulation results, wherein the simulation results include risk probability distribution, risk root cause chain, estimated cost and estimated construction period.

[0044] Specifically, step S2 includes the following steps: S21: For each candidate scheduling scheme, extract the qualification and timeliness data of the corresponding resources and the information flow association data between tasks from the project spatiotemporal attribute map.

[0045] In this embodiment, the system traverses a given candidate scheduling scheme, which specifies which resource(s) will execute each task in which time window.

[0046] First, for each personnel resource involved in the plan, the system queries and extracts its "qualification validity data" from the "Project Spatiotemporal Attribute Map" based on its unique identifier. This data is a list, and each record in the list corresponds to a qualification, including at least the qualification type (such as "Military Industry Level 2 Confidentiality Qualification" or "Senior Software Designer"), qualification level, and the qualification's effective and expiration dates. The system pays particular attention to the relative relationship between the qualification expiration date and the time window in which the resource is assigned to perform tasks.

[0047] Secondly, the system extracts "information flow association data" from the graph. This requires pre-modeling the information transfer or dependency relationships between tasks in the graph; for example, the output document of task A is an input dependency of task B. During implementation, a directed graph can be constructed based on the project's WBS decomposition and design document, where nodes represent tasks and edges represent information flows or strong dependencies. Edges can be attached with contextual attributes, such as the type or security level of the information being transferred. In this step, the system analyzes the time sequence of task scheduling in the current candidate solutions, combined with this directed information flow graph, to determine which tasks, upon completion, might have their output information flow to which subsequent tasks, thus forming potential information propagation paths. These two types of data (qualification validity and information flow association) are the core inputs for subsequent dynamic risk simulation.

[0048] S22: Construct the risk simulation model based on the qualification timeliness data and the information flow association data, wherein the risk simulation model is used to simulate the probability of qualification failure events occurring along the time axis and the impact of the task interruption it triggers, and to simulate the risk of information spreading across tasks along the path defined by the information flow association data.

[0049] In this embodiment, those skilled in the art can construct a hybrid model based on discrete event simulation and probabilistic graphical model as a risk extrapolation model. The model uses the timeline of the data extracted in S21 and the candidate solutions as a blueprint. For the "qualification failure risk," the model defines a failure probability distribution for each resource near the expiration date of each qualification (e.g., using a log-normal distribution to simulate the uncertainty of failure due to renewal delays). When the simulation clock advances to these time points, the model will randomly determine whether the qualification "fails" based on probability. If it fails, it immediately determines whether the resource is currently executing a task that depends on this qualification; if so, it triggers a "task interruption" event. For the "information diffusion risk," the model defines an unauthorized or unexpected "diffusion" probability for each information propagation path based on information flow association. This probability can be set based on factors such as task confidentiality level differences and historical behavioral data of the executors (e.g., whether violations have occurred).

[0050] In the simulation, once information is generated from the source task, the model follows the information flow path and randomly determines, with probability, whether the information will be improperly disseminated to downstream unauthorized task contexts. The model records the type, time of occurrence, and involved tasks and resources of all such risk events, thus simulating how a single risk event (such as a qualification failure) can trigger task interruptions, personnel rescheduling, and potential chain reactions, as well as the compliance risks that improper information dissemination may cause.

[0051] S23: Run the risk simulation model to perform multiple full-cycle simulations. In each simulation, trigger random risk events based on the probability model and record the resulting task impact sequence, cumulative additional costs, and cumulative project delays.

[0052] In this embodiment, the risk simulation model constructed by S22 is run using the Monte Carlo simulation method. A large number of simulations N is set (e.g., 5000 or 10000). Each simulation starts a new, independent simulation instance from the project start time. The simulation clock advances in the smallest time unit (e.g., day) or uses an event-based scheduling method. In each simulation run, whether and when all risk events (qualification failure, information spread) are triggered is determined independently and randomly by their preset probability model, reflecting the uncertainty of risk occurrence. The model needs to simulate the response logic after a risk occurs. For example, when a task is interrupted due to the executor's qualification failure, the model needs to simulate an emergency scheduling process (e.g., activating backup personnel). This process will generate additional "personnel switching costs" and "delays caused by task restart and connection". The system will fully record all triggered risk events in this simulation run, the response measures caused by the events, and the additional costs (e.g., additional manpower costs, penalties for breach of contract) and project delays incurred by these measures. The output of a single simulation run is a set of "task impact sequences" (recording that task A is interrupted at time T due to risk X, and is subsequently taken over by resource Y, etc.), a "cumulative additional cost" value, and a "cumulative project delay" value. Through a large number of independent simulations, a probability distribution of the possible project outcomes under this candidate scheduling scheme can be obtained.

[0053] S24: Perform causal aggregation analysis on the task impact sequence recorded in multiple simulations to generate the risk root cause chain that reflects the risk transmission path; and statistically analyze the cumulative additional costs and cumulative project delays in multiple simulations to generate the risk probability distribution, the estimated cost, and the estimated project duration.

[0054] In this embodiment, the "task impact sequence" from N simulated runs is first aggregated and analyzed. The system employs data mining methods, such as association rule analysis or causal discovery algorithms, to identify frequently occurring risk transmission patterns from massive event sequences. For example, the analysis reveals that "resource R's qualification B fails" has an 80% probability of causing "task T to be interrupted," and "task T's interruption" has a 60% probability of further causing its subsequent task, "task U's delayed start," to be delayed. Linking these high-frequency causal relationships in a logical order forms one or more "risk root cause chains," which intuitively reveal how the initial risk source gradually amplifies and ultimately affects the project's key objectives.

[0055] Secondly, statistical analysis is performed on the two values ​​of "cumulative additional cost" and "cumulative project delay" output from N simulations. By sorting all cost values ​​from smallest to largest, a cumulative distribution function can be generated, thus obtaining a "risk probability distribution" such as "the probability that the total project cost does not exceed the budget C is P%". Typically, "estimated cost" is defined as a certain statistic of these cost values, such as the median or expected value; and "estimated project duration" is defined as the expected value of the original planned project duration plus the project delay.

[0056] Finally, for a candidate scheduling scheme, this step outputs its quantitative risk profile, including the probability distribution and expected value of cost and schedule, and most importantly, the qualitative risk transmission mechanism (risk root cause chain).

[0057] S3: Input the risk probability distribution, estimated cost and estimated construction period corresponding to each candidate scheduling scheme into the multi-objective optimization model for solution and trade-off, and generate the optimal scheduling scheme.

[0058] Specifically, step S3 includes the following steps: S31: For each candidate scheduling scheme, extract the comprehensive risk index value from the risk probability distribution, and use it together with the estimated cost and the estimated construction period as input parameters of the multi-objective optimization model.

[0059] In this embodiment, the probability-based "risk profile" output in step S2 needs to be transformed into a quantitative indicator that can be used for deterministic or multi-objective comparisons. First, for each candidate scheduling scheme, the system extracts one or more "comprehensive risk indicator values" from its corresponding "risk probability distribution" (usually represented as a cumulative distribution function of the total project cost or total duration). A typical indicator is the "cost overrun probability," which is the probability value calculated for the final total project cost to exceed a preset budget threshold.

[0060] Another commonly used metric is "Value at Risk" (VaR) or "Conditional Value at Risk" (CVaR), which represents the additional costs or delays a project may face at a given confidence level. For example, one could calculate "the maximum possible additional costs to the project at a 95% confidence level."

[0061] In this way, a complete probability distribution is condensed into one or more representative scalar values. Then, this "comprehensive risk index value" (e.g., cost overrun probability P) is combined with the "estimated cost" C (e.g., expected cost) and "estimated duration" D (e.g., expected duration) obtained from the statistics in S24, forming an input vector [P, C, D] characterizing the multidimensional performance of the scheme. This vector will serve as the basic data input for the multi-objective optimization model to evaluate and compare various schemes.

[0062] S32: Set the weight relationship between the risk objective function, cost objective function and schedule objective function, and calculate based on the input parameters and the weight relationship to output the comprehensive evaluation value of each candidate scheduling scheme.

[0063] In this embodiment, the multi-objective optimization model needs to establish explicit objective functions to quantify the merits of different solutions. The model defines three core objective functions: the risk objective function FR (corresponding to the comprehensive risk index P, aiming to minimize it), the cost objective function FC (corresponding to the estimated cost C, aiming to minimize it), and the schedule objective function FT (corresponding to the estimated schedule D, aiming to minimize it).

[0064] To weigh multiple conflicting objectives, those skilled in the art can use a linear weighted sum method to transform them into a single-objective comprehensive evaluation value. In implementation, a set of weighting coefficients is pre-defined by the project manager or based on the project phase strategy. , , ( These represent the relative importance placed on risk, cost, and schedule, respectively. For example, in the early stages of a project, risk control might be given more emphasis. (Higher), in the final stage, the construction period may be more important ( (Higher). Next, for each candidate scheduling scheme's input vector... First, normalize each dimension (to eliminate the influence of dimensions), then calculate their weighted sum: comprehensive evaluation value. ,in , , These are the normalized values. Through this calculation, each scheme obtains a "comprehensive evaluation value" that can be used for total order comparison. The smaller the value, the better the overall performance of the solution under the current weight preference.

[0065] S33: Based on the comprehensive evaluation value, sort all candidate scheduling schemes, select schemes according to a set ratio, and form a Pareto front solution set.

[0066] In this embodiment, one of the purposes of calculating the comprehensive evaluation value is to efficiently select the set of truly excellent and comparable solutions, namely the Pareto front solution set.

[0067] First, the system sorts all candidate scheduling schemes in ascending order based on the "comprehensive evaluation value" calculated by S32 (the lower the score, the better).

[0068] Then, the system does not simply select the top-ranked solution, because the optimal solution under a single weight might overlook other potentially good solutions. The system will select a group of solutions with better "overall evaluation values" from the ranking list according to a preset proportion (e.g., the top 20%) to form an "elite solution pool".

[0069] Next, the system uses Pareto dominance to precisely filter this elite pool: it iterates through each solution in the pool, determining if there exists another solution that is superior in at least one of the three objectives (risk, cost, and schedule) and not inferior to it in all other objectives. If a solution is dominated by other solutions, it is eliminated. The remaining solutions, which are mutually independent, constitute the "Pareto front solution set." The solutions in this set represent the boundary where existing solutions can no longer simultaneously improve all objectives; each solution is the optimal solution under a specific trade-off among the three objectives.

[0070] S34: Perform conflict target analysis on the schemes in the Pareto front solution set, and select the final optimal scheduling scheme from the Pareto front solution set according to the preset decision preference.

[0071] In this embodiment, a final selection needs to be made from multiple non-inferior Pareto optimal solutions.

[0072] First, the system performs visualization or quantitative analysis on the Pareto front solution set to reveal the conflicting relationships between objectives. For example, it can generate a two-dimensional or three-dimensional scatter plot to show the distribution of each solution in the solution set across dimensions such as risk-cost and cost-time, clearly telling decision-makers "how much risk needs to be reduced, and how much cost or time may need to be increased."

[0073] Then, the system automatically or assistedly selects the final solution based on a preset, more refined "decision preference" rule. This preference rule can be an automated decision-making strategy, such as: "In the Pareto front, select the lowest cost among all solutions whose duration does not exceed the contract period D_max"; or "Select the shortest duration among all solutions whose risk probability is below the threshold R_max". If the preference rule cannot uniquely determine a solution, the system can submit several solutions that meet the initial screening criteria, along with their detailed multi-dimensional data (including their risk root cause chains), to the project manager for final manual decision-making. The selected solution is output as the "optimal scheduling solution" generated by the entire scheduling method in the current cycle, and is used to enter the execution phase.

[0074] S4: During the execution of the optimal scheduling scheme, real-time status data is collected as new evidence and input into the risk inference model to update the risk probability distribution, and a rescheduling signal is triggered when the preset conditions are not met.

[0075] Specifically, step S4 includes the following steps: S41: During the execution of the optimal scheduling scheme, real-time status data reflecting task progress, resource status and external events are collected according to the set sampling frequency.

[0076] In this embodiment, after the "optimal scheduling scheme" enters the execution phase, the system starts a background monitoring service to actively collect multi-dimensional real-time status data according to a preset sampling frequency (e.g., once every 4 hours, or at fixed times every day).

[0077] In practice, the system integrates with existing enterprise project management systems, human resource systems, access control / log systems, etc., via APIs to automatically acquire data. The collected data mainly includes three categories: Task progress data: Synchronize the actual start time, actual completion percentage, current blocking issues, consumed man-hours, and costs of each task from the project management system.

[0078] Resource status data: Synchronize the on-duty / off-duty status (such as leave, business trip) and current workload of personnel from the human resources system; synchronize the latest status of all personnel's qualification certificates from the qualification management system, especially the approval progress of qualification renewal applications, which is used to dynamically update the uncertainty of qualification expiration.

[0079] External event data: Capture external events that may affect the project, such as changes in customer requirements, critical equipment failures, and policy and regulatory updates, by integrating with the enterprise notification system or by manual entry.

[0080] All collected raw data will be timestamped and converted into a uniform format to form time-stamped "evidence" records, providing input for subsequent real-time analysis.

[0081] S42: Input the real-time status data as time-series evidence into the risk inference model, drive the risk inference model to perform probability inference, and output the updated risk probability distribution.

[0082] In this embodiment, the system uses the timestamped "real-time status data" collected in step S41 as new observational evidence and inputs it into the "risk inference model" constructed for the current execution plan in step S2. This model is constructed in step S22 as a hybrid model based on a probabilistic graphical model (such as a dynamic Bayesian network). During implementation, the system performs a probabilistic inference process: the model fixes these new pieces of evidence (e.g., "Personnel A's XX qualification renewal application has been submitted for approval today," "Task B's actual completion rate is 2 days behind schedule") onto the corresponding observation variable nodes in the network. Then, it calls a probabilistic graphical model inference algorithm (such as a confidence propagation algorithm) to recalculate the posterior probability distribution of all unknown variables in the network based on this new deterministic evidence. This means that the model will dynamically increase the probability of subsequent task delays based on the fact that "the task has been delayed"; or appropriately decrease the probability of the qualification expiring in the near future based on the evidence that "qualification renewal has entered the process." After the inference calculation is completed, the model outputs an updated "risk probability distribution" reflecting the latest current situation. This distribution is more accurate than the initial predicted distribution because it incorporates the actual state of the project so far.

[0083] S43: Compare the updated risk probability distribution with the preset risk probability thresholds, and calculate the progress deviation between the actual project progress and the estimated construction period.

[0084] In this embodiment, the system performs two parallel quantitative assessments. First, a risk threshold comparison is performed: Before the project starts, the manager sets warning thresholds and action thresholds for key risk indicators. For example, the warning threshold for the "probability of project interruption due to qualification issues" is set at 20%, and the action threshold is set at 40%. The system reads the "updated risk probability distribution" output by S42, extracts the current probability values ​​of key risk indicators (such as the probability of project total cost overrun and the probability of core task chain failure), and compares them with the preset thresholds.

[0085] Secondly, schedule variability is calculated: The system summarizes and calculates the current overall actual progress of the project from the "task progress data" collected by S41 (e.g., based on the critical path completion rate or the proportion of total man-hours consumed). Simultaneously, it obtains the planned progress (i.e., the planned value on the "estimated duration" curve) at the same point in time from the currently executing "optimal scheduling plan." Schedule variability (SV) is typically calculated as the difference between the actual progress and the planned progress; it can also be calculated as a time-related deviation, such as the number of days the actual completion date is delayed compared to the planned date. These two assessment results (various risk probability values ​​and schedule variability values) will be quantified and standardized for subsequent automated decision-making.

[0086] S44: When it is detected that the probability value of any risk in the updated risk probability distribution exceeds its corresponding risk probability threshold, or the progress deviation exceeds the preset tolerance range, it is determined that the preset condition is not met, and the rescheduling signal is triggered.

[0087] In this embodiment, the system implements automatic decision-making logic to determine whether rescheduling needs to be initiated. This logic is based on two evaluation results in S43. The system presets two types of "preset conditions": 1) the probability values ​​of each risk are all below their corresponding acceptable thresholds (usually action thresholds); 2) the project schedule deviation is within the allowable tolerance range (e.g., schedule delays do not exceed 5 days). The system checks these two conditions in real time. The "OR" logic means that if either condition is broken, it is considered "not meeting the preset conditions". For example, even if the schedule is normal, but the probability of "information security risk" spikes to 45% (exceeding the 40% action threshold) due to the detection of abnormal data access logs, the system will also determine it as abnormal. Or, even if the probabilities of each risk are controllable, but the schedule deviation has reached 7 days (exceeding the 5-day tolerance), it is also determined as abnormal. Once it is determined that the preset conditions are not met, the system immediately generates a structured "rescheduling signal". This signal is an event object that contains at least the following information: trigger type (risk exceeding threshold / schedule exceeding limit), the specific risk indicator or deviation value triggered, the trigger time, and the associated current "risk root cause chain" as a diagnostic reference. This signal will be sent to the event handling center, thereby automatically triggering the rescheduling process in step S5.

[0088] S5: In response to the rescheduling signal, a rescheduling process is triggered based on real-time status data, the updated risk probability distribution, and the associated risk root cause chain.

[0089] Specifically, step S5 includes the following steps: S51: In response to the rescheduling signal, obtain a snapshot of the project site at the current moment.

[0090] In this embodiment, when the system receives the structured "rescheduling signal" from step S44, it immediately triggers the data snapshot acquisition process to obtain a complete and consistent instantaneous view of the project status at the decision moment, i.e., a "project site snapshot". In specific implementation, the system calls a monitoring interface similar to that in step S41, but this acquisition is immediate and one-time, aiming to freeze the project status at the moment the "rescheduling signal" is triggered.

[0091] This snapshot integrates at least the following core data: 1) Progress snapshot: Obtain the precise latest status of all incomplete tasks from the project management system, including the percentage of completion, actual start time, current blocking issues, and calculate the total cost and total man-hours actually consumed since the start of the project.

[0092] 2) Real-time resource status: Obtain the latest on-the-job status, current tasks and progress of all personnel from the human resources system, and synchronize the real-time validity status of all qualification certificates from the qualification management system (especially the current stage of the qualification renewal process that is under review).

[0093] 3) Event Log: This section integrates historical logs of all recorded risk events, external events, and responses since the last scheduling or rescheduling. All this data is tagged with a unified timestamp (i.e., the rescheduling trigger time) and encapsulated into a structured data object. This "project snapshot" replaces the initial blueprint during scheduling, serving as the sole true and reliable starting point for the rescheduling process.

[0094] S52: Using the project site snapshot as the new initial state, and combining the updated risk probability distribution and the risk root cause chain, adjust the project task requirements and constraints.

[0095] In this embodiment, the system uses the "project site snapshot" obtained in S51 as a benchmark to intelligently adjust future "project task requirements" and "constraints," rather than simply repeating the initial plan. This is the key to achieving "adaptability" in rescheduling.

[0096] First, the system redefines the planning scope: tasks that have been fully completed in the snapshot are marked as history and will no longer be scheduled; tasks that are currently being executed, with their remaining unfinished workload, are reassessed as new "subtasks"; tasks that have not yet started remain unchanged. This constitutes a new set of tasks.

[0097] Secondly, adjustments are made by integrating risk intelligence: the system deeply analyzes the "updated risk probability distribution" and associated "risk root cause chains" that triggered this rescheduling. For example, if the root cause chain shows that "personnel A's X qualification is about to expire" is the main risk, the system will add a strong constraint of "must confirm successful qualification renewal" to the new constraints for tasks that this person may participate in, or directly exclude them from the candidate resources of recent high-risk tasks. If the risk distribution shows an increased risk of information leakage of a certain type, the system may adjust the confidentiality level of subsequent related tasks or forcibly add information isolation requirements. At the same time, based on the actual progress deviation, the system will dynamically adjust the total duration constraints or cost budget of the remaining work. Finally, a set of updated "project task requirements" and "constraints" injected with risk intelligence for the remaining project cycle is output.

[0098] S53: Based on the adjusted project task requirements and constraints, and the pre-built project spatiotemporal attribute map, multiple candidate scheduling schemes are regenerated to trigger and complete the rescheduling process.

[0099] In this embodiment, the system takes the adjusted "project task requirements" and "constraints" output by S52 as input and reuses the "pre-built project spatiotemporal attribute map" (the resource status in this map has been updated through snapshots) to re-execute the complete intelligent scheduling calculation. Specifically, this process is an iterative cycle from steps S1 to S3, but the starting state and input conditions are updated. First, based on the new tasks and constraints, and combined with the latest resource availability data (obtained from snapshots), the system runs the S1 process to generate multiple "candidate scheduling schemes" for the remaining part of the project. Next, for these new schemes, the S2 process is executed to build a risk simulation model and perform a full-cycle risk simulation, quantifying their risks, costs, and schedules under the new circumstances. Then, the S3 process is executed to select a new "optimal scheduling scheme" from the new schemes through a multi-objective optimization model. This new scheme is the optimal arrangement for the project execution phase after the current point in time. The system will automatically replace the old schemes that have been interrupted or are no longer applicable with this new scheme and instruct each execution system (such as the project management system) to proceed according to the new plan. At this point, the "rescheduling process" has been completed, and the project will continue to be executed under the guidance of a new, more realistic, and pre-emptively risk-avoiding optimal scheduling scheme.

[0100] Example 2 This invention also provides an intelligent project scheduling system based on engineering management, used to execute the aforementioned intelligent project scheduling method based on engineering management, with reference to... Figure 2 As shown, the scheduling system includes: The candidate scheme generation module 100 is used to generate multiple candidate scheduling schemes based on a pre-built spatiotemporal attribute map of the project and the input project task requirements and constraints.

[0101] The cycle risk simulation module 200 is used to construct a risk simulation model for each candidate scheduling scheme and perform a full cycle risk simulation, and output the corresponding simulation results, wherein the simulation results include risk probability distribution, risk root cause chain, estimated cost and estimated construction period.

[0102] The multi-objective optimization decision module 300 is used to input the risk probability distribution, the estimated cost and the estimated construction period corresponding to each candidate scheduling scheme into the multi-objective optimization model for solution and trade-off, so as to generate the optimal scheduling scheme.

[0103] The rescheduling triggering module 400 is used to collect real-time status data as new evidence to input into the risk inference model during the execution of the optimal scheduling scheme, update the risk probability distribution, and trigger a rescheduling signal when the preset conditions are not met.

[0104] The rescheduling execution module 500 is used to respond to the rescheduling signal and trigger the rescheduling process based on real-time status data, the updated risk probability distribution, and the related risk root cause chain.

[0105] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0106] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, including read-only memory (ROM), random access memory (RAM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), one-time programmable read-only memory (OTPROM), electrically erasable programmable read-only memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disc storage, disk storage, magnetic tape storage, or any other computer-readable medium capable of carrying or storing data.

[0107] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

Claims

1. An intelligent project scheduling method based on engineering management, characterized in that, The scheduling method includes the following steps: Based on the pre-built spatiotemporal attribute map of the project, as well as the input project task requirements and constraints, multiple candidate scheduling schemes are generated. For each candidate scheduling scheme, a risk simulation model is constructed and a full-cycle risk simulation is performed, and the corresponding simulation results are output. The simulation results include the risk probability distribution, the risk root cause chain, the estimated cost, and the estimated construction period. The risk probability distribution, estimated cost, and estimated duration corresponding to each candidate scheduling scheme are input into a multi-objective optimization model for solution and trade-off, thereby generating the optimal scheduling scheme. During the execution of the optimal scheduling scheme, real-time status data is collected as new evidence and input into the risk inference model to update the risk probability distribution, and a rescheduling signal is triggered when the preset conditions are not met. In response to the rescheduling signal, a rescheduling process is triggered based on real-time status data, the updated risk probability distribution, and the associated risk root cause chain.

2. The intelligent project scheduling method based on engineering management according to claim 1, characterized in that, The risks include at least one of the following: compliance risks, continuity risks, and information security risks.

3. The intelligent project scheduling method based on engineering management according to claim 2, characterized in that, Based on the pre-built project spatiotemporal attribute map and the input project task requirements and constraints, multiple candidate scheduling schemes are generated, including: Obtain the resource status data from the project's spatiotemporal attribute map, the task attribute data from the project's task requirements, and the constraints. Based on the resource status data and the task attribute data, a multi-dimensional matching calculation is performed to generate an initial scheduling scheme set. The schemes in the initial scheduling scheme set are verified for compliance and feasibility based on the constraints, resulting in a verified scheme set. Based on a preset scheduling strategy, the schemes in the verified scheme set are adjusted and optimized to output the multiple candidate scheduling schemes.

4. The intelligent project scheduling method based on engineering management according to claim 3, characterized in that, The preset scheduling strategy is at least one of the following: total cost minimization strategy, resource load balancing strategy, and task priority guarantee strategy.

5. The intelligent project scheduling method based on engineering management according to claim 4, characterized in that, For each candidate scheduling scheme, a risk simulation model is constructed and a full-cycle risk simulation is performed, outputting the corresponding simulation results, including: For each candidate scheduling scheme, extract the corresponding resource qualification and timeliness data and the information flow association data between tasks from the project spatiotemporal attribute map; The risk simulation model is constructed based on the qualification timeliness data and the information flow association data. The risk simulation model is used to simulate the probability of qualification failure events occurring along the time axis and the impact of the task interruption it triggers, and to simulate the risk of information spreading across tasks along the path defined by the information flow association data. The risk simulation model is run multiple times for full-cycle simulation. In each simulation, random risk events are triggered based on the probability model, and the resulting task impact sequence, cumulative additional costs, and cumulative project delays are recorded. A causal aggregation analysis is performed on the task impact sequence recorded in multiple simulations to generate the risk root cause chain that reflects the risk transmission path; and the cumulative additional costs and cumulative project delays from multiple simulations are statistically analyzed to generate the risk probability distribution, the estimated cost, and the estimated project duration.

6. The intelligent project scheduling method based on engineering management according to claim 5, characterized in that, The step of inputting the risk probability distribution, estimated cost, and estimated construction period corresponding to each candidate scheduling scheme into a multi-objective optimization model for solution and trade-offs includes: For each candidate scheduling scheme, a comprehensive risk index value is extracted from the risk probability distribution and used together with the estimated cost and the estimated construction period as input parameters of the multi-objective optimization model. The weighting relationships between the risk objective function, cost objective function, and schedule objective function are set, and the comprehensive evaluation value of each candidate scheduling scheme is output based on the input parameters and the weighting relationships.

7. The intelligent project scheduling method based on engineering management according to claim 6, characterized in that, The generation of the optimal scheduling scheme includes: All candidate scheduling schemes are sorted based on the comprehensive evaluation value, and schemes are selected according to a set ratio to form a Pareto front solution set. Conflict target analysis is performed on the schemes in the Pareto front solution set, and the final optimal scheduling scheme is selected from the Pareto front solution set according to the preset decision preferences.

8. The intelligent project scheduling method based on engineering management according to claim 7, characterized in that, During the execution of the optimal scheduling scheme, real-time status data is collected as new evidence and input into the risk inference model to update the risk probability distribution. A rescheduling signal is triggered when preset conditions are not met. This includes: During the execution of the optimal scheduling scheme, real-time status data reflecting task progress, resource status and external events are collected at a set sampling frequency. The real-time status data is used as time-series evidence and input into the risk inference model to drive the risk inference model to perform probability inference and output the updated risk probability distribution. The updated risk probability distribution is compared with the preset risk probability thresholds, and the progress deviation between the actual project progress and the estimated construction period is calculated. When the probability value of any risk in the updated risk probability distribution exceeds its corresponding risk probability threshold, or the progress deviation exceeds the preset tolerance range, it is determined that the preset conditions are not met, and the rescheduling signal is triggered.

9. The intelligent project scheduling method based on engineering management according to claim 8, characterized in that, In response to the rescheduling signal, based on real-time status data, the updated risk probability distribution, and the related risk root cause chain, a rescheduling process is triggered, including: In response to the rescheduling signal, obtain a snapshot of the project site at the current moment; Using the project site snapshot as the new initial state, and combining the updated risk probability distribution and the risk root cause chain, adjust the project task requirements and constraints. Based on the adjusted project task requirements and constraints, as well as the pre-constructed project spatiotemporal attribute map, multiple candidate scheduling schemes are regenerated to trigger and complete the rescheduling process.

10. An intelligent project scheduling system based on engineering management, used to execute the intelligent project scheduling method based on engineering management as described in any one of claims 1 to 9, characterized in that, The scheduling system includes: The candidate solution generation module is used to generate multiple candidate scheduling solutions based on a pre-built spatiotemporal attribute map of the project and the input project task requirements and constraints. The cycle risk simulation module is used to construct a risk simulation model for each candidate scheduling scheme and perform a full cycle risk simulation, and output the corresponding simulation results, wherein the simulation results include risk probability distribution, risk root cause chain, estimated cost and estimated construction period; The multi-objective optimization decision module is used to input the risk probability distribution, the estimated cost and the estimated construction period corresponding to each candidate scheduling scheme into the multi-objective optimization model for solution and trade-off, and generate the optimal scheduling scheme. The rescheduling trigger module is used to collect real-time status data as new evidence to input into the risk inference model during the execution of the optimal scheduling scheme, update the risk probability distribution, and trigger a rescheduling signal when the preset conditions are not met. The rescheduling execution module is used to respond to the rescheduling signal and trigger the rescheduling process based on real-time status data, the updated risk probability distribution, and the related risk root cause chain.

Images