A construction engineering whole cycle collaborative management and risk prediction system

Abstract

The application relates to the technical field of building engineering information management, and discloses a building engineering whole-cycle collaborative management and risk prediction system, which comprises a resource token definition module, a process transaction modeling module, a discrete event pre-performance simulation engine and a deadlock detection module. The resource token definition module maps physical elements into discrete resource tokens which are configured with unit time lag rate. The process transaction modeling module encapsulates a process into an atomic transaction unit containing request and release instructions. The discrete event pre-performance simulation engine executes flow transfer on a virtual time axis and records the suspension duration. The deadlock detection module monitors the occupation topology in real time to identify a cyclic waiting closed loop. A dynamic priority arbitration logic calculates a cumulative lag weight value according to the product of the rate and the duration, and forcibly allocates resources according to the cumulative lag weight value to eliminate the deadlock. The application converts physical lag cost into a calculation weight by constructing a dynamic arbitration algorithm based on a time value gradient, and improves the ability of the system to automatically converge to an optimal solution under complex constraints.

Description

Technical Field

[0001] This invention relates to a collaborative management and risk prediction system for the entire lifecycle of construction projects, belonging to the field of information management technology for construction projects. Background Technology

[0002] Currently, mainstream engineering management data processing systems generally use the critical path method or program review and approval technique to construct schedule models. These models typically abstract complex engineering projects into static time-linked lists based on the logical relationships between work processes, focusing on handling the time constraints between processes. Based on this, they calculate the theoretical shortest duration and critical nodes of the project. However, in scenarios where multiple processes are executed concurrently in deep foundation pits or high-rise building construction with severely limited resources, the aforementioned static time-linked list-based data models have inherent defects due to the lack of physical constraints. The models are built on the premise of unlimited resource supply or static allocation settings, and cannot identify or calculate dynamic blockages caused by instantaneous competition for discrete resources such as construction site space, large machinery, or working capital. When multiple processes simultaneously request the same scarce resource in physical space and time, even if the time logic of each process does not conflict, the competitive deadlock of the resource topology can lead to a complete halt in the actual workflow. The logical connection and resource status become disconnected, allowing the system to only record delayed events and lacking the ability to anticipate systemic paralysis caused by resource competition. Traditional statistical regression prediction based on historical data is insufficient to handle transient resource deadlock logic on site, and relying solely on manual experience for on-site scheduling cannot cope with the explosive challenges of massive concurrent combinations of work processes.

[0003] Existing technologies attempt to introduce dynamic control methods, but they are mostly limited to numerical balance based on efficiency indicators and lack the ability to deconstruct deep logical loops. For example, Chinese invention patent CN113298430B discloses an AI and BIM-based IPD risk collaborative design management system for building engineering projects. It dynamically adjusts the number of construction workers by calculating the degree of risk impact. However, based on the efficiency difference feedback adjustment mechanism, it essentially deals with the quantity of resources rather than their order. When deadlocks occur on site due to spatiotemporal exclusion or machine monopoly, i.e., when multiple processes are waiting in a loop, simply increasing or decreasing the number of workers cannot break the logical deadlock. On the contrary, it exacerbates the on-site congestion due to the oversaturation of resources. The system lacks a deterministic arbitration mechanism that identifies and forcibly blocks deadlocks from the resource topology level.

[0004] Therefore, how to construct a mechanism for the deterministic inference of spatiotemporal logic and resource status that deeply integrates data processing systems, and how to identify and prevent logical deadlocks in advance in multi-transaction concurrent environments, has become the technical problem to be solved by this invention. Summary of the Invention

[0005] To address the problems mentioned in the background art, the technical solution of the present invention is as follows: A collaborative management and risk prediction system for the entire life cycle of construction projects, the system comprising:

[0006] The resource token definition module is used to map physical workspace, machinery and equipment and funding amount into discrete resource tokens, and to configure a unit time lag rate for each resource token. This unit time lag rate is used as a static parameter to characterize the time value weight of the corresponding resource when it is occupied.

[0007] The process transaction modeling module is used to encapsulate work processes into atomic transaction units containing resource token request instructions and resource token release instructions. The atomic transaction units are configured with logic settings to enter a suspended waiting state when the requested resource token cannot be obtained.

[0008] The discrete event pre-simulation engine is used to execute the flow logic of atomic transaction units on a virtual timeline and record the simulation clock duration of each atomic transaction unit in the suspended waiting state in real time.

[0009] The deadlock detection and blocking module is used to monitor the resource token occupancy topology in real time during the operation of the discrete event pre-simulation engine. When multiple atomic transaction units are detected to have formed a closed-loop stagnation state due to circular waiting for resource tokens, the dynamic priority arbitration logic is triggered. The dynamic priority arbitration logic executes the following operation rules: it traverses all atomic transaction units involved in the closed-loop stagnation state, calculates the cumulative lag weight based on the product of the unit time lag rate and the simulation clock duration, uses the cumulative lag weight as the dynamic sorting basis, forcibly allocates the requested resource token to the atomic transaction unit with the highest cumulative lag weight, releases the circular waiting state, and generates the corresponding scheduling instruction.

[0010] Preferably, the resource token definition module divides the construction site into multiple logical grids and defines the space occupancy right of each logical grid as an exclusive resource token; when constructing atomic transaction units, the process transaction modeling module transforms spatial location constraints into request logic for exclusive resource tokens. This request logic enforces that when atomic transaction units request exclusive resource tokens, they must simultaneously satisfy the availability of the time window and the adjacency constraint of the logical grid, ensuring the continuity and executability of the generated scheduling instructions in the physical space topology.

[0011] Preferably, the process transaction modeling module is configured with a resource release mechanism that follows a two-way handshake handover logic for atomic transaction units to ensure the authenticity of the resource token state transition. The specific execution logic is as follows: when the first atomic transaction unit, acting as the upstream, issues a resource token release instruction, the system does not directly release the resource token, but marks it as an intermediate state pending handover and logically locks it to the predetermined downstream second atomic transaction unit; only after receiving a confirmation signal from the second atomic transaction unit, the state of the resource token is updated to occupied; if a rejection signal is received, the state rollback logic is triggered, the resource token release instruction of the first atomic transaction unit is revoked and reset to the incomplete state.

[0012] Preferably, the deadlock detection and blocking module uses a resource allocation graph algorithm to construct a real-time resource allocation matrix. The real-time resource allocation matrix is ​​arranged with atomic transaction units as rows and resource tokens as columns. The closed-loop stagnation state is determined by detecting whether there is a closed loop in the directed graph corresponding to the real-time resource allocation matrix. The discrete event pre-simulation engine is equipped with a fast-forward simulation mechanism, which is used to complete multiple iterative calculations of the flow logic of all atomic transaction units within the time step of the actual construction cycle, thereby realizing the advance prediction of potential deadlock risks.

[0013] Preferably, the calculation rule for the cumulative lagging option weight in the dynamic priority arbitration logic is configured as follows: For each target atomic transaction unit i involved in the closed-loop stagnation state, the system calculates its corresponding dynamic weight W using the following formula. i : Among them, W i R is the cumulative hysteresis weight of the target atomic transaction unit i, N is the total number of resource tokens currently held by this transaction unit, and R is the cumulative hysteresis weight of the target atomic transaction unit i. d,k T is the unit time delay rate for the kth held resource token. hold,k The time the k-th resource token has been held under the simulation clock; the system is based on the calculated W. i The values ​​are sorted in descending order for all target atomic transaction units, with priority given to satisfying W. i The resource request of the atomic transaction unit with the largest value.

[0014] Preferably, the resource token definition module is also used to define funding resource tokens and set total thresholds and dynamic consumption rules for these funding resource tokens; the process transaction modeling module decomposes and maps the project budget into the amount of funding tokens required for the execution of atomic transaction units; during the simulation process, the discrete event pre-simulation engine verifies in real time whether the remaining amount of funding tokens in the current system meets the request of the atomic transaction unit to be activated. If it does not meet the requirement, the atomic transaction unit is forced to enter a suspended waiting state until other atomic transaction units release enough funding tokens, thereby realizing the constraint and pre-simulation of the risk of project cash flow interruption at the simulation level.

[0015] Preferably, the instruction distribution and execution module includes a differential feedback adjustment unit. This unit is used to compare the simulation completion time after deadlock resolution with the preset baseline plan time after the discrete event pre-simulation engine is released from the closed-loop stagnation state. When the deviation between the two exceeds a preset threshold, the high-frequency scarce resource tokens that caused the deadlock are automatically identified based on the records generated by the dynamic priority arbitration logic, and a resource replenishment suggestion instruction to increase the amount of such resource tokens is generated, or a logic optimization instruction to adjust the execution sequence of atomic transaction units on non-critical paths is generated.

[0016] Preferably, the discrete event pre-simulation engine also includes an adaptive duration correction module based on historical credit, which is used to count the frequency of each atomic transaction unit triggering state rollback logic within the historical running cycle and calculate the corresponding rework probability coefficient; when executing future simulation pre-simulations, the engine automatically extends the estimated execution time of the corresponding type of atomic transaction unit according to the rework probability coefficient, and introduces quantitative compensation for the uncertainty of the construction period caused by construction quality risks in the simulation model.

[0017] Preferably, the system is deployed in a distributed computing architecture, with the resource token definition module, the process transaction modeling module, and the deadlock detection and blocking module running as independent service nodes; the state data of the resource token is stored in a strongly consistent distributed database to ensure the atomicity and uniqueness of the resource token's occupancy and release states in the global view when multiple terminals concurrently request resource tokens, preventing resource overselling or state conflicts caused by data read and write delays.

[0018] Preferably, the instruction distribution and execution module is also equipped with a visualization mapping unit, which is used to map the optimal execution sequence containing deadlock resolution paths generated by the discrete event pre-simulation engine into a Gantt chart or a multi-dimensional construction simulation animation. The visualization mapping unit is specially configured to highlight the time nodes that triggered the dynamic priority arbitration logic during the simulation process and the atomic transaction units involved, intuitively displaying potential risk deadlock points and the arbitration resolution path recommended by the system.

[0019] Compared with the prior art, the beneficial effects of the present invention are:

[0020] 1. In the collaborative management and risk prediction of the entire construction project cycle, a resource token definition module and a process transaction modeling module are constructed. Discrete physical work spaces, machinery and equipment and funding amounts are mapped to resource tokens with total thresholds and counting attributes. Work processes are reconstructed into atomic transaction units with strict token input and output rules. The atomic transaction units are forced to activate only after simultaneously capturing a set of resource tokens that meet the preset logical rules. Loose business logic is transformed into a strict data transaction protocol. Logical connections and strong coupling of resource status are established at the data layer to eliminate scheduling conflicts caused by resource competition and ensure the physical logic executability of scheduling instructions and the dynamic balance of resource topology.

[0021] 2. Utilizing a discrete event pre-simulation engine and deadlock detection module, the flow logic of atomic transaction units is traversed in fast-forward mode on the virtual timeline. A real-time resource allocation matrix is ​​constructed using a resource allocation graph algorithm to detect whether there are closed loops representing circular waiting in the matrix. Future logical deadlock states are identified in advance at the data logic layer. Based on a deterministic inference calculation mechanism, potential stalled nodes are accurately located before physical conflicts occur, and timing adjustment strategies are automatically triggered to resolve closed loops, thereby realizing the transformation of risk control from delayed response to endogenous immunity.

[0022] 3. Configure a unit time delay rate attribute in the resource token data structure. During the simulation process, calculate the total accumulated delay value of atomic transaction units in the blocked or held state in real time. The deadlock detection and blocking module uses the instantaneous growth rate of the accumulated value as a dynamic weight to calculate the release benefit ratio of each competing transaction unit. Based on this, execute the priority release or state rollback decision, and transform the resource holding cost over time into a computable data weight. This ensures that the system automatically tends to the solution with the lowest total logical cost in the deadlock resolution scheduling operation, realizing the underlying unification of scheduling logic and resource utilization efficiency. Embed a two-way handshake delivery and state rollback protocol in the atomic transaction unit flow logic. Define the resource token release process as an intermediate state to be delivered, and force downstream atomic transaction units to confirm the availability logic of the received resources. Once a rejection signal is received, the state rollback logic is immediately triggered, the upstream transaction completion state is revoked and the token holding right is reset. Utilize the logical dependency relationship between upstream and downstream transactions to build an endogenous data verification closed loop. Without relying on additional sensing hardware, the self-correction capability of the process logic eliminates the accumulation of false state information, ensuring the authenticity of system state data and the accuracy of scheduling decisions. Attached Figure Description

[0023] Figure 1 This is a flowchart of the system data processing based on resource token transfer and deadlock arbitration of the present invention;

[0024] Figure 2 This is a comparison chart of the transaction unit suspension time distribution and optimization under the current priority arbitration mechanism of this invention;

[0025] Figure 3 This is a diagram of the closed-loop architecture of virtual simulation and real-time feedback in the cloud, which is the basis of this invention. Detailed Implementation

[0026] The present invention will be described in detail below with reference to specific embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0027] This invention provides a collaborative management and risk prediction system for the entire lifecycle of construction projects. The system includes a resource token definition module, a process transaction modeling module, a discrete event pre-simulation engine, a deadlock detection and blocking module, and an instruction distribution and execution module. The resource token definition module maps the physical workspace, machinery, and funding amounts at the construction site into discrete resource tokens. This module configures a unit time delay rate for each resource token. The static parameter representing the time value weight of the corresponding resource when it is occupied is used. In specific implementation, the system divides the construction site into multiple logical grids and defines the space occupancy right of each logical grid as an exclusive resource token. The process transaction modeling module encapsulates the work process into atomic transaction units containing resource token request instructions and resource token release instructions. The atomic transaction unit is configured with a logic setting to enter a suspended waiting state when the requested resource token cannot be obtained. When constructing the atomic transaction unit, the spatial location constraint is transformed into the request logic for the exclusive resource token. The atomic transaction unit is forced to satisfy the availability of the time window and the adjacency constraint of the logical grid when requesting the exclusive resource token.

[0028] The discrete event pre-simulation engine executes the flow logic of atomic transaction units on a virtual timeline and records the simulation clock duration of each atomic transaction unit in the suspended waiting state in real time. The engine is equipped with a fast-forward simulation mechanism to perform multiple iterative calculations of the flow logic of all atomic transaction units within a preset future period, within the actual construction cycle's time step. The deadlock detection and blocking module monitors the resource token occupancy topology in real time during the discrete event pre-simulation engine's operation. This module uses a resource allocation graph algorithm to construct a real-time resource allocation matrix, with atomic transaction units as rows and resource tokens as columns. By detecting whether a closed loop exists in the directed graph corresponding to this matrix, it determines whether a closed-loop stagnation state has formed. When multiple atomic transaction units are detected to have formed a closed-loop stagnation state due to cyclically waiting for resource tokens, dynamic priority arbitration logic is triggered. The dynamic priority arbitration logic executes the following calculation rules: iterates through all target atomic transaction units involved in the closed-loop stagnation state, and calculates the cumulative lag weight value based on the product of the unit time lag rate and the simulation clock duration. The specific calculation formula is as follows: Among them, W i R is the cumulative hysteresis weight of the target atomic transaction unit i, N is the total number of resource tokens currently held by this transaction unit, and R is the cumulative hysteresis weight of the target atomic transaction unit i. d,k T is the unit time delay rate for the kth held resource token. hold,k Given the held time of the k-th resource token under the simulation clock, the system calculates W. i The values ​​are sorted in descending order for all target atomic transaction units, forcing them to W. i The atomic transaction unit with the highest value allocates the resource token it requested to release the circular wait state and generate the corresponding scheduling instruction.

[0029] The atomic transaction unit is configured with a resource release mechanism that follows a two-way handshake settlement logic. When the first atomic transaction unit, acting as the upstream, issues a resource token release command, the system marks the resource token as an intermediate state awaiting settlement and logically locks it to a predetermined downstream second atomic transaction unit. Only after receiving an acknowledgment signal from the second atomic transaction unit does the system update the resource token's status to occupied. If a rejection signal is received, the system triggers a state rollback logic, canceling the first atomic transaction unit's resource token release command and resetting it to an incomplete state. Regarding the unit time delay rate R in the resource token definition module... d The quantitative setting is not a preset fixed value, but is determined through regression calculation based on historical financial data. It reads the settlement list database of similar completed projects, extracts records of work stoppages caused by specific resource shortages, sums up equipment rental fees, labor idle time costs, and contract penalties, and divides the result by the cumulative work stoppage time to obtain the benchmark rate. The benchmark rate is then multiplied by the real-time price index of the current construction area, and the product is the final R. d The value is written to the resource token attribute field, establishing a direct calculation path from physical cost data to algorithm weight parameters, thus accumulating the lag weight value W. i The calculation results correspond to the actual economic losses of the engineering entity, without relying on manual experience estimation; the simulation time step Δt of the discrete event pre-simulation engine and the trigger threshold of the differentiated feedback adjustment unit are locked based on numerical convergence test and statistical distribution characteristics. The initial step size is set as the greatest common divisor of the standard working hours of all atomic transaction units, and a step size halving comparison operation is performed; when the deviation of two consecutive total project duration prediction values ​​is less than 0.1%, the current step size is determined as the discretization interval, the standard deviation σ of the daily progress completion rate of the historical construction log is calculated, and the activation threshold of the differentiated feedback adjustment unit is set to 3σ; the logic filters random noise fluctuations within the confidence interval and limits the triggering conditions of resource replenishment suggestion instructions to statistically significant systematic blocking events.

[0030] Example 1: In a foundation pit project of a large commercial complex, three parallel processes are involved: earthwork removal, concrete internal support erection, and deep well dewatering. The project is limited by a single on-site trestle ramp, defined as resource token A; and a tower crane covering the work surface, defined as resource token B. The earthwork removal process, i.e., atomic transaction unit 1, needs to continuously occupy resource token A and intermittently request resource token B; the internal support erection process, i.e., atomic transaction unit 2, needs to occupy resource token B and intermittently request resource token A as a temporary storage area. When the discrete event pre-simulation engine advances to day 20 on the virtual timeline, the deadlock detection and blocking module identifies that atomic transaction unit 1 holds resource token A and requests resource token B, while atomic transaction unit 2 holds resource token B and requests resource token A, forming a closed-loop stagnation state of circular waiting. The dynamic priority arbitration logic is triggered, and the preset unit time delay rate R is read. dAmong them, the downtime fee rate R for the earthmoving truck fleet associated with atomic transaction unit 1 d,1 The idle rate R for the self-owned hoisting equipment associated with Atomic Transaction Unit 2 is set at 3000 yuan / hour. d,2 The price is set at 400 yuan per hour, and the system uses the formula... Calculations show that the cumulative lag weight W1 of atomic transaction unit 1 is higher than the cumulative lag weight W2 of atomic transaction unit 2. Based on this, the system generates a forced scheduling instruction, triggering the state rollback logic of atomic transaction unit 2 to release the resource token B it holds, and then allocates resource token B to atomic transaction unit 1. After executing the scheduling instruction, atomic transaction unit 1 captures resource token B, completes the task, and releases resource tokens A and B. Atomic transaction unit 2 is then reactivated and acquires resources to hoist the steel beam, restoring the continuity of the internal support erection process. This process shows that the system uses a dynamic arbitration algorithm based on time value gradients to transform multi-process competition into a deterministic weight comparison under the physical constraints of limited resources, thereby resolving the deadlock state by minimizing the total logical cost of the project.

[0031] Example 2: In a simulation verification experiment based on real high-rise building construction data, this experiment built a discrete event pre-simulation platform deployed on a high-performance distributed computing cluster, loaded an engineering topology model containing 520 atomic transaction units and 45 heterogeneous resource tokens. The experimental data came from the entire construction log and financial settlement list of a completed 30-story Grade A office building project. In the experimental parameter setting procedure, the simulation time step Δt was determined to be 10 minutes based on the derivation of the Nyquist sampling theorem in discrete event systems; the unit time delay rate R... d Based on the equipment rental unit price and labor subcontracting hourly rate in the project's commercial contract, and through regression analysis of historical risk data, the risk coefficient was determined. The numerical values; this experiment designs a multi-dimensional control system, and sets a unified original input benchmark: on the 35th day of the foundation construction phase, a random disturbance event is injected into the system, that is, the core resource token No. 1 tower crane is set to be unavailable for the next 72 hours due to a fault, and Gaussian white noise with a standard deviation of 15% is superimposed on the execution time of all processes. Based on this benchmark, three groups of test objects are set: control group 1 adopts the critical path method (CPM) logic with only static time linked list constraints; control group 2 removes the dynamic priority arbitration logic and adopts the first-come, first-served (FIFO) linear allocation strategy. The complete operation of the sample group of this invention includes deadlock detection and the weighting of options based on cumulative lag W. i The arbitration logic.

[0032] After the experiment started, the discrete event pre-simulation engine advanced the virtual clock. When the simulation reached the fault injection point on day 35, control group 1 failed to detect the missing resource tokens. Subsequent processes requested resources as planned, leading to an accumulation of invalid resource request queues and passive delays in the actual project schedule. Control group 2, although recognizing resource scarcity, allocated resources to atomic transaction unit X, which was on a non-critical path but had an earlier request time, during the competition for alternative resources. This forced atomic transaction unit Y, which was on the critical path and held expensive leased equipment, to suspend. After detecting the deadlock risk, the sample group of this invention calculated and compared the weights of the competing units: W of atomic transaction unit X... X =400, while the W of the atomic transaction unit Y Y =2500, see Table 1, the system prioritizes allocating resources to atomic transaction units Y with higher W values.

[0033] Table 1: Comparison of key process data and final performance of different experimental groups in dealing with resource deadlock.

[0034]

[0035] Data shows that the prototype of this invention minimized project delays and reduced total logic costs by approximately 39% compared to control group 2. Further gradient stress testing indicates that the cost advantage of the prototype of this invention expands as the duration of the fault increases, especially when resource scarcity leads to W... i When the calculated values ​​generally exceed the preset threshold, the system performance curve shows an inflection point, automatically triggering a resource replenishment suggestion command.

[0036] Example 3: This example combines Figures 1 to 3 This document describes a collaborative management and risk prediction system for the entire lifecycle of a construction project. Figure 1 As shown, the data processing flow begins at the physical element input stage, where the workspace, machinery, and funding amount are input into the resource token definition module. After physical mapping and setting the unit time lag rate, the data is transmitted to the process transaction modeling module via data mapping. Here, atomic transaction encapsulation is completed, bidirectional handshake handover logic is set, and the transaction unit is injected. The system then enters the discrete event pre-simulation engine, where the process flows on the virtual timeline and the suspension duration is recorded. The resulting real-time state stream is transmitted to the deadlock detection and blocking module for resource allocation topology monitoring and closed-loop identification. Once a deadlock is identified, dynamic priority arbitration logic is triggered. By calculating the cumulative lag weight and forcing resource allocation, a deadlock-resolving solution is generated. Finally, the instruction distribution and execution module generates the optimal scheduling instruction or performs a visualization mapping. This module is connected back to the simulation engine through a normal scheduling feedback loop and also includes a visualization mapping unit to generate a Gantt chart and highlight deadlock points.

[0037] like Figure 2As shown, the horizontal axis represents the pending wait time intervals in 10-minute increments, and the vertical axis represents the number of transaction units. The legend distinguishes between the horizontal bar chart without arbitration logic and the diagonal bar chart with dynamic priority arbitration. The data shows that without arbitration logic, a large number of transaction units are concentrated in the medium-to-high waiting time range of 20 to 60 minutes. However, after introducing dynamic priority arbitration, the peak number of transaction units shifts significantly to the left to the low waiting time range of 0 to 20 minutes, and the number is extremely small in the long waiting time range of more than 50 minutes. Figure 3 As shown, the architecture is divided into a virtual simulation cloud layer and a physical construction site layer. The cloud layer performs fast-forward iterative calculations on the virtual timeline and deploys a discrete event simulation engine as the core computing node, a deadlock arbitration and detection service responsible for dynamic weight calculation, and a distributed state database. The cloud layer receives data from the physical layer through resource tokenization mapping and issues scheduling instructions downwards. The physical construction site layer provides real-time execution feedback on the real timeline, including a mobile handheld terminal for initiating work process requests, equipment positioning sensors, and a large on-site visualization screen for displaying deadlock risk warnings. There is a two-way handshake confirmation mechanism between the mobile handheld terminal and the large on-site visualization screen, thus forming a closed-loop control system for virtual-real interaction.

[0038] Example 4: In the interior decoration construction of super high-rise buildings, the project faces the challenge of high-density spatiotemporal overlapping operations, especially in the standard floor corridors. Ceiling electromechanical pipeline installation, wall dry-hanging stone keel welding, and floor flammable material laying need to be carried out in parallel within the same time window. Traditional progress management models can only identify physical spatial overlap conflicts, but cannot identify the adjacency safety risks caused by the mutual exclusion of process attributes. This example aims to provide a transparent execution logic with clear algorithm paths and parameter setting procedures for logical grid adjacency constraints and logical optimization instruction generation mechanisms. The system divides the standard floor corridor into a 3-meter by 3-meter discrete logical grid through the resource token definition module, and defines each grid as a space resource token S with exclusive attributes. x,y When constructing atomic transaction units, the process transaction modeling module not only encapsulates the request instruction for the target mesh space token, but also writes the repulsion radius parameter r according to the physical properties of the process. ex Define the keel welding process as atomic transaction unit A, configure it with hot work attribute, and set r ex,A =1; Define the carpet laying process as atomic transaction unit B, configured with flammable properties, and set r ex,B =0, when the discrete event pre-simulation engine runs to 10:00 AM on the 120th day, atomic transaction unit A requests to occupy grid S. 5,5 Welding operations are being performed, while atomic transaction unit B already holds the adjacent grid S. 5,6 Carpets are laid.

[0039] The deadlock detection and blocking module executes an extended detection algorithm based on shadow tokens. The system receives an atomic transaction unit A on S... 5,5 When making a request, based on its r ex,A Parameters, not only checking S 5,5 The occupancy status is also automatically monitored, and a set of virtual shadow requests is automatically generated to lock the S... 5,5 The system detects the safety attribute bits of all adjacent meshes centered at Chebyshev distance 1. 5,6 The system has determined that a security logic deadlock exists because the atomic transaction unit A is already occupied by atomic transaction unit B and its attribute is flammable. Therefore, the activation request of atomic transaction unit A is rejected, and it is forcibly placed into a suspended waiting state. To resolve this deadlock, dynamic priority arbitration logic intervenes in the calculation, and the system extracts the time delay rate R of the work group and equipment unit associated with atomic transaction unit A. d,A The price is 800 yuan per hour, and the waiting time T has been set. hold,A The unit time delay rate R for atomic transaction unit B is 0.5 hours. d,B The price is 200 yuan per hour, and the holding period is T. hold,B It takes 2 hours, according to the formula. Calculate and obtain W A =400, and W B =400. Under the boundary condition of equal weights, the differential feedback adjustment unit in the instruction dispatch and execution module triggers the constraint relaxation strategy. This strategy does not directly interrupt any process, but executes the space swap search algorithm: the system traverses other unoccupied logical grids of the same type on the same floor and finds grid S 20 meters away from the current position. 10,5 If the resource is idle and its neighborhood satisfies security constraints, the system generates a space redirection instruction to move the target resource of atomic transaction unit A from S. 5,5 Change to S 10,5 After the instruction was issued, the welding team was relocated to the other end of the corridor to carry out the work, while the carpet laying team continued to work in the original position. The two teams maintained a safe distance from each other in physical space, and there was no substantial delay in the construction period.

[0040] Example 5: In a large airport terminal construction project with extremely high requirements for parameter calibration accuracy, this example constructs an offline calibration and data filling procedure. During the initial system deployment phase, this procedure connects to the project's historical financial database and equipment rental contract management system via a data interface module. It extracts actual settlement data from three previously completed similar projects. Based on a data cleaning algorithm, the system removes outliers caused by force majeure. Then, it performs statistical regression analysis on the unit time cost of each type of resource (such as tower cranes, mobile scaffolding, and specialized welding equipment) to calculate a comprehensive rate benchmark value including basic rental costs, labor costs, and potential default risk penalties. The system executes a series of Monte Carlo simulation experiments, setting different R... d Input values ​​into the simulation engine and observe the backtesting effect of the system-generated scheduling instructions in historical scenarios. Specifically, compare the deviations of the simulated total duration and total cost with the actual historical values. After multiple iterative optimizations, the system automatically locks in the R-value that minimizes the backtesting deviation. d The value is used as the formal calibration parameter for this type of resource and written into the system's parameter configuration database.

[0041] To address potential environmental differences and system cold start issues during the initial deployment phase, this embodiment also implements a pre-deployment calibration procedure. Before the system officially goes live, a one-week shadow operation test is conducted. This involves the system receiving resource requests and status update data from the field in real time and running a complete set of deadlock detection and arbitration logic in the background without issuing actual scheduling commands. Instead, the generated commands are compared and analyzed with the actual manual scheduling results on-site. The system focuses on monitoring the false alarm rate and false negative rate of logical deadlocks, as well as the consistency between the arbitration results and the decisions of experienced on-site schedulers. When the system frequently triggers false deadlocks for a specific process, it automatically triggers an adaptive calibration algorithm to fine-tune the exclusive time window threshold or adjacency exclusion radius parameter of the resource tokens associated with that process until the scheduling suggestions output by the system match the optimal decisions made on-site to the preset acceptance criteria.

[0042] Example 6: To address the differences in location, scale, and management model among various construction projects, this example constructs an offline generation and verification process for an adaptive parameter matrix. Through the engineering feature extraction module, key feature vectors are parsed from the BIM model and construction organization design documents of the project to be deployed, including building height, single-story area, process concurrency, and resource constraint index. The system uses these feature vectors to perform similarity retrieval in a pre-set engineering knowledge graph, matches the closest historical benchmark project case, and extracts the verified initial parameter set as the cold start parameter matrix for the project to be deployed.

[0043] During the fine-tuning and calibration phase of the parameter matrix, the system executes an optimization procedure based on a multi-objective genetic algorithm. This procedure uses the cold-start parameter matrix as the initial value of the population and sets the schedule deviation rate, resource utilization rate, and logic deadlock frequency as multi-dimensional fitness functions. The system performs no less than 500 generations of population iterative evolution in an offline simulation environment. Each generation of individuals undergoes full-cycle simulation through a discrete event pre-simulation engine. During the iteration process, the system introduces a dynamic mutation operator to simulate random disturbances such as extreme weather and sudden equipment failures to test the robustness of the parameter matrix under non-ideal conditions. Finally, the system outputs the parameter combination in the Pareto optimal solution set as the official operating parameters for the project and generates a calibration report that includes parameter sensitivity analysis, clearly indicating the tolerance range and adjustment basis of each key parameter.

[0044] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0045] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A construction project whole cycle collaborative management and risk prediction system, characterized in that, The system comprises: a resource token definition module, configured to map the physical working space, mechanical equipment and capital amount into discrete resource tokens, and to configure a unit time penalty rate for each resource token as a static parameter representing the time value weight of the corresponding resource when occupied; a process transaction modeling module, configured to encapsulate the working process into an atomic transaction unit containing resource token request instructions and resource token release instructions, and to configure the atomic transaction unit with a logic setting of entering a suspended waiting state when the requested resource token cannot be acquired; a discrete event pre-play simulation engine, configured to execute the flow logic of the atomic transaction unit on a virtual time axis, and to record the simulation clock duration of each atomic transaction unit in the suspended waiting state in real time; a deadlock detection and blocking module, configured to monitor the occupation topology of the resource token in real time during the running of the discrete event pre-play simulation engine, and to trigger a dynamic priority arbitration logic when a closed loop stagnation state is detected due to the circular waiting of resource tokens by multiple atomic transaction units; wherein the dynamic priority arbitration logic performs the following operation rules: traversing all atomic transaction units involved in the closed loop stagnation state, calculating the cumulative penalty weight value according to the product of the unit time penalty rate and the simulation clock duration, taking the cumulative penalty weight value as the dynamic sorting basis, and forcibly allocating the requested resource token to the atomic transaction unit with the highest cumulative penalty weight value to remove the circular waiting state and generate the corresponding scheduling instruction.

2. The construction project whole cycle collaborative management and risk prediction system according to claim 1, characterized in that, The resource token definition module divides the construction site into multiple logical grids, and defines the space occupation right of each logical grid as an exclusive resource token; the process transaction modeling module converts the spatial position constraint into a request logic for the exclusive resource token when constructing the atomic transaction unit, which forcibly constrains the atomic transaction unit to meet the availability of the time window and the adjacency constraint of the logical grid when requesting the exclusive resource token.

3. The construction project whole cycle collaborative management and risk prediction system according to claim 1, characterized in that, The process transaction modeling module configures the atomic transaction unit with a resource release mechanism following a two-way handshake delivery logic to ensure the authenticity of the resource token state flow, and the specific execution logic is as follows: when the first atomic transaction unit as the upstream issues a resource token release instruction, the system does not directly release the resource token, but marks it as a pending delivery intermediate state and logically locks it to the predetermined second atomic transaction unit downstream; only after receiving the confirmation signal from the second atomic transaction unit, the state of the resource token is updated to occupied; if a rejection signal is received, the state rollback logic is triggered to cancel the resource token release instruction of the first atomic transaction unit and reset it to an incomplete state.

4. The construction project whole cycle collaborative management and risk prediction system according to claim 1, characterized in that, The deadlock detection and blocking module uses the resource allocation graph algorithm to construct a real-time resource allocation matrix, which takes the atomic transaction unit as the row and the resource token as the column, and determines the closed loop stagnation state by detecting whether there is a closed loop in the corresponding directed graph of the real-time resource allocation matrix; The discrete event pre-play simulation engine is configured with a fast-forward simulation mechanism to complete multiple iterations of the flow logic of all atomic transaction units in the future preset period within the time step of the actual construction period.

5. The construction project whole cycle collaborative management and risk prediction system according to claim 1, characterized in that, In the dynamic priority arbitration logic, the calculation rule for the cumulative lagging weight is configured as follows: For each target atomic transaction unit i involved in the closed-loop stagnation state, the system calculates its corresponding dynamic weight W using the following formula. i : Among them, W i R is the cumulative hysteresis weight of the target atomic transaction unit i, N is the total number of resource tokens currently held by this transaction unit, and R is the cumulative hysteresis weight of the target atomic transaction unit i. d,k T is the unit time delay rate for the kth held resource token. hold,k The time the k-th resource token has been held under the simulation clock; the system is based on the calculated W. i The values ​​are sorted in descending order for all target atomic transaction units, with priority given to satisfying W. i The resource request of the atomic transaction unit with the largest value.

6. The construction project whole cycle collaborative management and risk prediction system according to claim 1, characterized in that, The resource token definition module is further configured to define a fund type resource token, and set a total amount threshold and a dynamic consumption rule for the fund type resource token; the procedure transaction modeling module decomposes and maps the engineering budget into an atomic transaction unit to execute a required fund token request amount; The discrete event pre-play simulation engine verifies in real time whether the fund token remaining amount of the current system meets the request of the atomic transaction unit to be activated in the simulation process, and if not, forces the atomic transaction unit to enter a suspended waiting state until other atomic transaction units release sufficient fund tokens, thereby realizing constraint and pre-play of the engineering cash flow disruption risk at the simulation level.

7. The construction project whole cycle collaborative management and risk prediction system according to claim 1, characterized in that, The instruction distribution and execution module includes a differential feedback adjustment unit configured to compare the simulation completion time after the deadlock is resolved with a preset benchmark plan time after the discrete event pre-play simulation engine is released from the closed-loop stall state; When the deviation between the two exceeds a preset threshold, the high-frequency scarce resource token causing the deadlock is automatically identified based on the record generated by the dynamic priority arbitration logic, and a resource supply suggestion instruction to increase the amount of the high-frequency scarce resource token is generated or a logical optimization instruction to adjust the execution timing of the atomic transaction unit on a non-critical path is generated.

8. The construction project whole cycle collaborative management and risk prediction system according to claim 3, characterized in that, The discrete event pre-play simulation engine further includes an adaptive duration correction module based on historical credit, configured to count the frequency of state rollback logic triggered by each atomic transaction unit in a historical operation period, and calculate a corresponding rework probability coefficient; during future simulation pre-play, the engine automatically extends the estimated execution duration of the corresponding type of atomic transaction unit according to the rework probability coefficient, and introduces quantitative compensation for the construction quality risk caused by the uncertainty of the construction period in the simulation model. 9.The building engineering whole cycle collaborative management and risk prediction system according to claim 1, characterized in that, The system is deployed in a distributed computing architecture, and the resource token definition module, the procedure transaction modeling module, and the deadlock detection and blocking module are operated as independent service nodes.

10. The construction project whole cycle collaborative management and risk prediction system according to claim 1, characterized in that, The instruction distribution and execution module is further configured with a visual mapping unit configured to map the optimal execution timing of the discrete event pre-play simulation engine containing the deadlock resolution path into a Gantt chart or a multi-dimensional construction simulation animation; the visual mapping unit is specially configured to highlight the time nodes and the atomic transaction units involved in the simulation process that have triggered the dynamic priority arbitration logic.

Images