AI-based road construction resource optimization configuration method and system

Abstract

The present application relates to the technical field of resource scheduling, in particular to a road construction resource optimization configuration method and system based on AI, comprising the following steps: analyzing the operation area and time period, judging the spatial overlap and time concurrency between nodes, identifying resource code conflicts, calculating node priority ranking scores and reconstructing connection structures, extracting path indicators to calculate delay contribution and risk distribution, screening abnormal equipment to calculate adaptive scores, and adjusting the scheduling order to generate scheduling arrangement results. In the present application, by constructing the task conflict intensity index and the priority ranking score mechanism, combining the delay contribution ratio and the path risk structure, quantifying the equipment performance using the scheduling adaptation level score, and adjusting and optimizing the instruction arrangement structure using the resource call order, the dynamic reorganization of the construction task scheduling chain and the delay risk suppression of the task path are realized, the equipment matching efficiency and the stability of task execution are improved, and the response capability of resource allocation and the self-adaptive control capability of path scheduling are enhanced.

Description

Technical Field

[0001] This invention relates to the field of resource scheduling technology, and in particular to an AI-based method and system for optimizing the allocation of road construction resources. Background Technology

[0002] Resource scheduling technology encompasses the rational allocation and management of limited resources across time and space to ensure the efficient execution of various tasks or projects. It focuses on task allocation, time planning, resource coordination, and scheduling strategy optimization, and is widely applied in various industries such as manufacturing, transportation, engineering construction, and information processing. In road construction scenarios, resource scheduling focuses on the coordinated management of construction equipment, personnel, materials, and construction processes. Using parameters such as time windows, construction areas, and task priorities, resources are dynamically allocated and planned to improve the overall operational efficiency and responsiveness of construction organization. Among these, AI-based road construction resource optimization allocation methods utilize artificial intelligence models and algorithms to dynamically schedule and optimize the resources involved in road construction. Addressing the challenges of diverse construction resources, strong temporal dependencies, and complex scheduling scheme combinations, this approach involves constructing machine learning models for resource status prediction, introducing graph-based task relationship modeling methods, and employing deep reinforcement learning for task and resource matching strategy optimization. By combining specific factors such as the task dependency sequence of different processes in the road construction workflow, equipment availability constraints, and worker skill characteristics, an executable construction scheduling scheme is formed, automating the resource allocation and time arrangement process.

[0003] Existing road construction resource optimization technologies focus on static management of resource coordination and time allocation, lacking a quantitative expression of scheduling conflicts caused by spatial relationships and temporal overlaps of tasks. This makes it difficult to identify potential bottleneck chains between tasks when multiple construction areas intersect or resources are reused, and fails to fully integrate task topology characteristics and path dependencies. The lack of a dynamic feedback mechanism between task sequencing and equipment allocation makes it difficult to identify and intervene in advance when task delay nodes are delayed. For example, when the same equipment is called multiple times in multiple paths but the response cycle is unstable, it is impossible to accurately judge the intensity of scheduling risk, resulting in resource stacking, scheduling imbalance, and cascading delays of subsequent tasks. Summary of the Invention

[0004] The purpose of this invention is to address the shortcomings of existing technologies by proposing an AI-based method and system for optimizing the allocation of road construction resources.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: an AI-based road construction resource optimization allocation method, comprising the following steps:

[0006] S1: Call the road construction task information, analyze the work area boundary and work period of the task node, determine the spatial overlap and temporal concurrency between adjacent nodes, compare the resource codes of the task equipment configuration items, determine the conflict intensity of the task node, and generate the task node topology.

[0007] S2: Based on the task node topology, analyze the task phase sequence, upstream dependencies and spatial density, calculate the task node priority ranking score, rearrange the node order according to the task number, reconstruct the connection structure of the task graph, and generate scheduling node priority data.

[0008] S3: Call the priority data of the scheduling node, extract the duration, attendance stability, equipment waiting time and interference conditions of the path node, calculate the delay contribution ratio, determine the intersection node and delay risk distribution, and generate path delay impact index;

[0009] S4: Based on the path delay impact index, analyze the correspondence between the equipment number and the task process type involved in the affected path, determine the equipment response time and execution error deviation, screen abnormal equipment, calculate the scheduling adaptation level score, and generate scheduling adaptation level data.

[0010] S5: Based on the scheduling adaptation level data, analyze the device task frequency and idle time, determine the scheduling interval, compare the matching degree between the call order and the adaptation level, adjust the device scheduling order, update the arrangement structure of resources in the scheduling instruction table, and generate the resource scheduling arrangement result.

[0011] As a further aspect of the present invention, the task node topology includes spatial intersection relationship type, time-overlapping task group number, and resource code repetition mapping relationship; the scheduling node priority data specifically includes stage sequence number, path depth identifier, and spatial distribution density; the path delay impact index includes delay contribution ratio distribution, node intersection location index, and delay risk associated path number; the scheduling adaptation level data specifically includes equipment task matching frequency, scheduling response time label, and job error offset amplitude; and the resource scheduling arrangement result includes resource number sorting position, update order in the instruction table, and equipment scheduling priority identifier.

[0012] As a further aspect of the present invention, the step of obtaining the task node topology structure specifically includes:

[0013] S111: Obtain road construction task information, analyze the work area boundary information and work time plan of task nodes, determine the area overlap and work time coverage between adjacent nodes, and obtain task overlap combination.

[0014] S112: Call the task cross-over combination, extract the equipment configuration information table of each node in the combination, compare the resource codes in the task equipment configuration items, identify resource configuration duplication, calculate the conflict intensity value of the task combination according to the area of ​​regional crossover, the proportion of overlapping operation time, and the proportion of resource code duplication, and generate the node conflict intensity distribution value by corresponding the conflict intensity value with the task node combination.

[0015] S113: Based on the node conflict intensity distribution value, analyze the conflict connection relationship between tasks, construct a conflict value weight matrix, establish undirected edge connection relationship according to the conflict value between nodes, label the conflict edge weight, insert the conflict edge into the initial task graph node set, establish the topological connection structure between nodes, and obtain the task node topological structure.

[0016] As a further aspect of the present invention, the step of obtaining the priority data of the scheduling node specifically includes:

[0017] S211: Obtain the topology of the task nodes, collect the stage sequence number and the number of upstream dependent nodes of each task node, determine the path depth of each node in the task dependency chain, and generate path depth distribution values.

[0018] S212: Based on the path depth distribution value, detect the node set of the region to which the task space belongs, compare the number of nodes and the space area of ​​each region, calculate the node density of the region, and associate the density data with the path depth distribution value to obtain the node density feature parameters;

[0019] S213: Call the node density feature parameters, calculate the task node priority ranking score based on the path depth distribution value, the number of upstream dependencies, and the density of the area where the node is located for each task node, adjust the connection relationship of the task nodes in the topology, reconstruct the node connection structure in the task graph, and generate scheduling node priority data.

[0020] As a further aspect of the present invention, the step of obtaining the path delay impact index specifically includes:

[0021] S311: Call the priority data of the scheduling node, extract the nodes in each task path, collect the duration of each node, personnel attendance stability, equipment scheduling waiting time and environmental interference conditions, analyze the actual delay impact of each node in this path, calculate the delay contribution ratio of the node, and generate the node delay contribution coefficient.

[0022] S312: Based on the node delay contribution coefficient, by judging the number of positions of the same node in multiple paths, intersecting nodes are filtered, and the path distribution characteristics of intersecting nodes are statistically analyzed to obtain the distribution parameters of intersecting nodes;

[0023] S313: Call the intersection node distribution parameters, determine the delay risk distribution structure of each intersection node under multiple paths, classify and identify the risk distribution density, output the risk level index of each intersection node, and generate the path delay impact index.

[0024] As a further aspect of the present invention, the step of obtaining the scheduling adaptation level data specifically includes:

[0025] S411: Based on the path delay impact index, analyze the task process type in each affected path, extract the equipment number associated with the corresponding task, compare the matching frequency of the equipment participating in the process with the process category required by the current task, and obtain process consistency label data.

[0026] S412: Call the process consistency tag data to determine the deviation of the equipment's response time in the scheduling start-up phase and the operation error in the execution phase, filter the equipment numbers with abnormal response time and operation error deviation, and establish a response deviation abnormality coefficient table;

[0027] S413: Based on the response offset anomaly coefficient table, extract the process matching frequency, response time and error offset rate of each device, calculate the scheduling adaptation level score of each device, and generate scheduling adaptation level data.

[0028] As a further aspect of the present invention, the step of obtaining the resource scheduling arrangement result specifically includes:

[0029] S511: Based on the scheduling adaptation level data, analyze the participation frequency of the currently assigned tasks for each device, detect the duration of unassigned tasks for each device, determine the scheduling interval period between consecutive tasks of the device, and generate device scheduling interval parameters.

[0030] S512: Call the device scheduling interval parameter, compare the matching degree between the device's scheduling position and the scheduling adaptation level in the current scheduling cycle, determine the consistency between the scheduling position and the adaptation level, adjust the arrangement order of the devices in the scheduling sorting table, and generate a device sorting adjustment sequence.

[0031] S513: Adjust the sequence according to the device sorting, update the arrangement structure of multiple device resources in the scheduling instruction table, integrate the sorting number of each device with the instruction mapping, and obtain the resource scheduling arrangement result.

[0032] An AI-based road construction resource optimization and allocation system, wherein the AI-based road construction resource optimization and allocation system is used to execute the aforementioned AI-based road construction resource optimization and allocation method, the system comprising:

[0033] The node conflict identification module calls up road construction task information, analyzes the work area boundary and work period of task nodes, determines the spatial overlap and temporal concurrency between adjacent nodes, compares the resource codes of task equipment configuration items, determines the conflict intensity of task nodes, and generates the task node topology.

[0034] The task priority ranking module analyzes the task stage order, upstream dependencies and spatial density based on the task node topology, calculates the task node priority ranking score, rearranges the node order according to the task number, reconstructs the connection structure of the task graph, and generates scheduling node priority data.

[0035] The path delay assessment module calls the priority data of the scheduling nodes, extracts the duration, attendance stability, equipment waiting time and interference conditions of the path nodes, calculates the delay contribution ratio, judges the intersection nodes and the distribution of delay risks, and generates path delay impact indicators.

[0036] The scheduling level determination module analyzes the correspondence between the equipment number and the task process type involved in the affected path based on the path delay impact index, determines the equipment response time and execution error deviation, filters abnormal equipment, calculates the scheduling adaptability level score, and generates scheduling adaptability level data.

[0037] The resource order adjustment module analyzes the device task frequency and idle time based on the scheduling adaptation level data, judges the scheduling interval, compares the matching degree between the call order and the adaptation level, adjusts the device scheduling order, updates the arrangement structure of resources in the scheduling instruction table, and generates the resource scheduling arrangement result.

[0038] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0039] In this invention, by constructing a task conflict intensity index and a priority ranking scoring mechanism, combined with the delay contribution ratio and path risk structure, the scheduling adaptation level scoring is used to quantify equipment performance, and the resource call ranking is used to adjust and optimize the instruction arrangement structure. This enables the dynamic reorganization of the construction task scheduling chain and the suppression of task path delay risks, improves equipment matching efficiency and task execution stability, enhances the responsiveness of resource allocation and the adaptive control capability of path scheduling, and optimizes the coordination and consistency between equipment, personnel and operation links in construction organization. Attached Figure Description

[0040] Figure 1 This is a schematic diagram of the workflow of the present invention;

[0041] Figure 2 This is a flowchart of the task node topology acquisition process of the present invention;

[0042] Figure 3This is a flowchart of the scheduling node priority data acquisition process of the present invention;

[0043] Figure 4 This is a flowchart of the path delay impact index acquisition process of the present invention;

[0044] Figure 5 This is a flowchart of the scheduling adaptation level data acquisition process of the present invention;

[0045] Figure 6 This is a flowchart of the process for obtaining the resource scheduling and arrangement results of the present invention. Detailed Implementation

[0046] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0047] In the description of this invention, it should be understood that the terms "length," "width," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientation or positional relationships, are based on the orientation or positional relationships shown in the accompanying drawings and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, in the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0048] Please see Figure 1 This invention provides a technical solution: an AI-based method for optimizing the allocation of road construction resources, comprising the following steps:

[0049] S1: Call the road construction task information, analyze the work area boundary and work period of the task node, determine the spatial overlap and temporal concurrency between adjacent nodes, compare the resource codes of the task equipment configuration items, determine the conflict intensity of the task node, and generate the task node topology.

[0050] S2: Based on the task node topology, analyze the task phase sequence, upstream dependencies and spatial density, calculate the task node priority ranking score, rearrange the node order according to the task number, reconstruct the connection structure of the task graph, and generate scheduling node priority data.

[0051] S3: Call the priority data of the scheduling node, extract the duration, attendance stability, equipment waiting time and interference conditions of the path node, calculate the delay contribution ratio, determine the intersection node and delay risk distribution, and generate path delay impact indicators.

[0052] S4: Based on the path delay impact index, analyze the correspondence between the equipment number and task process type involved in the affected path, determine the equipment response time and execution error deviation, screen abnormal equipment, calculate the scheduling adaptation level score, and generate scheduling adaptation level data.

[0053] S5: Based on the scheduling adaptation level data, analyze the device task frequency and idle time, determine the scheduling interval, compare the matching degree between the call order and the adaptation level, adjust the device scheduling order, update the arrangement structure of resources in the scheduling instruction table, and generate the resource scheduling arrangement result.

[0054] The task node topology includes spatial intersection relationship type, time-overlapping task group number, and resource code repetition mapping relationship. The scheduling node priority data specifically includes stage sequence number, path depth identifier, and spatial distribution density. The path delay impact indicators include delay contribution ratio distribution, node intersection location index, and delay risk associated path number. The scheduling adaptability level data specifically includes equipment task matching frequency, scheduling response time label, and operation error offset amplitude. The resource scheduling arrangement results include resource number sorting position, update order in the instruction table, and equipment scheduling priority identifier.

[0055] Please see Figure 2 The specific steps for obtaining the task node topology are as follows:

[0056] S111: Obtain road construction task information, analyze the work area boundary information and work time plan of task nodes, determine the area overlap and work time coverage between adjacent nodes, and obtain task overlap combination.

[0057] To obtain road construction task information, the system extracts all task entries numbered T01 to T50 from the task database. For each task entry, the system retrieves its "Work Area Boundary Information" field, which is stored as an array of polygon vertex coordinates. For example, the area for task T03 is [[121.45,31.22],[121.46,31.22],[121.46,31.21],[121.45,31.21]]. The system also retrieves the "Work Period Plan Content" field, which is stored as start and end timestamps. For example, the time period for T03 is [1765516800,1765545600]. Subsequently, the system... The system performs spatial geometric intersection operations between each pair of nodes to determine the overlap status of any two task polygon regions. For task pairs with overlapping regions, it further reads their work time periods to determine if there is an intersection between the two time intervals. For example, if the work areas of tasks T03 and T08 overlap by 1500 square meters, and the work time periods of T03 [1765516800, 1765545600] and T08 [1765527600, 1765556400] overlap by 18000 seconds, then the system stores the combination (T03, T08) in the initial conflict set. This process is repeated for all 50 task nodes. After several rounds of analysis, 112 pairs of task combinations that overlapped in both space and time were identified, forming task overlap combinations.

[0058] S112: Invoke the task overlap combination, extract the device configuration information table of each node in the combination, compare the resource codes in the task device configuration items, identify resource configuration duplication, and use the following formula based on the area of ​​regional overlap, the proportion of overlapping operation time, and the proportion of resource code duplication:

[0059] ;

[0060] Calculate the conflict intensity value of the task combination, map the conflict intensity value to the task node combination, and generate the node conflict intensity distribution value;

[0061] in, The conflict intensity value of the task combination is calculated by comprehensively considering parameters such as the area of ​​overlap between regions, the proportion of overlapping operation times, the repetition rate of resource allocation, the total number of tasks, and the task density. This is a normalized value for the area of ​​intersection between regions, obtained by dividing the actual area of ​​intersection between task pairs by the maximum area of ​​intersection among all tasks in the current project. This is a normalized value for the overlap ratio of task times. It is obtained by normalizing the ratio of the minimum overlap time of the two tasks to the minimum total time of the two tasks. The repetition ratio for task combination resources is determined by matching the device resource codes in the task pair and calculating the proportion of duplicate codes to the total number of resource codes. This is a normalized value representing the total number of tasks in the current task set, obtained by dividing the actual total number of tasks by a set task size reference value. The normalized task distribution density per unit area is obtained by normalizing the ratio of the average distribution density of the current task point to the reference maximum density.

[0062] The system calls for overlapping task combinations and processes one of the combinations (T03, T08). It extracts the "Device Configuration Information Tables" for T03 and T08 respectively. The resource codes for the configuration items in T03 are {E-001, E-002, C-005}, and the resource codes for the configuration items in T08 are {E-001, C-005, P-009}. Through set intersection operations, the system identifies the duplicate resource codes as {E-001, C-005}. Subsequently, it calculates the conflict strength value using the formula... The explanation and calculation process of each parameter are as follows. This formula quantifies the degree of conflict between specific task pairs by combining the direct conflict factors of space, time, and resources (numerator) with the overall macro-complexity factor of the project (denominator). The multiplication operation in the numerator represents the coupling effect of the three conflict factors, that is, the risk is amplified when all three factors exist simultaneously. The square root operation in the denominator is used to smooth the impact of the total number and density of tasks, avoiding excessive dilution of the conflict value due to the large scale of the project. Here, represents the conflict intensity value of the task combination, and represents the normalized value of the area of ​​intersection between regions. This is the normalized value of the overlap ratio of task times. Configure the repetition ratio for task combination resources. This is the normalized value of the total number of tasks in the current task set. To determine the normalized task distribution density per unit area, the specific assignment and calculation are as follows: First, by retrieving the intersection area of ​​all 112 overlapping combinations, the maximum intersection area of ​​all tasks in the current project is determined to be 5000 square meters. The actual intersection area between T03 and T08 is 1500 square meters. Secondly, the total operation time of T03 is 28,800 seconds, and the total operation time of T08 is 28,800 seconds. The minimum total operation time of the two is 28,800 seconds, and their overlap time is 18,000 seconds. Therefore, the time overlap ratio is... Normalize this ratio for all combinations, assuming 0.625 is the current maximum value. Furthermore, the union of the resource codes of T03 and T08 is {E-001, E-002, C-005, P-009}, the total number of resource codes is 4, and the number of duplicate codes is 2. Then, setting the task size reference value to 100, and the total number of tasks in the current task set to 50, then... The reference value of 100 is based on the average number of tasks (85) of the past 20 similar road construction projects, plus an increase of approximately 18% as a benchmark. The specific setting process is as follows: the average number of tasks in historical projects is obtained by statistical analysis. and standard deviation The reference value is set as follows Finally, by analyzing the geographic coordinates of all 50 tasks, the average distribution density of task points was calculated to be 4 task points per square kilometer, while the set maximum reference density was 10 task points per square kilometer. The reference maximum density value of 10 is based on historical data from high-density construction areas in the city. It analyzes the three most congested construction projects over the past five years, with peak project densities of 9.5, 10.2, and 10.8 projects per square kilometer, respectively. The average value is then rounded up. Substituting these values ​​into the formula:

[0063] ;

[0064] The conflict intensity value measures the comprehensive interference caused by spatial overlap, concurrent operation time, and repetitive resource allocation between any pair of task nodes during road construction. Essentially, it is a relative intensity coefficient based on a three-factor aggregation judgment. A larger value indicates a stronger scheduling conflict coupling risk between the task pair, requiring a tighter conflict edge connection in the task graph. This parameter serves as the basis for edge weights during graph structure construction, directly affecting the relative positions and connectivity between nodes in the graph, providing structural data for subsequent tasks such as task sequencing, priority evaluation, and graph reconstruction. The results show that the conflict intensity value for task pair (T03, T08) is 0.2096. The system repeats this calculation for all 112 combinations, mapping each combination to its calculated conflict intensity value, forming a key-value pair set of 112 entries, which represents the node conflict intensity distribution value. The advantage of the formula lies in the introduction of… and As a denominator adjustment factor, the calculation of conflict intensity is no longer limited to the micro-attributes of task pairs, but combines the overall macro-complexity of the project. When the overall project tasks are very dense, even if the direct spatiotemporal resource overlap between two tasks is not large, their potential conflict risk will be moderately amplified due to the tension of the overall environment. Conversely, in a sparse project environment, the conflict intensity calculated by the same direct overlap will be relatively reduced. This dynamic adjustment mechanism makes the conflict assessment results more in line with the complexity and urgency of the actual construction environment.

[0065] S113: Based on the node conflict intensity distribution value, analyze the conflict connection relationship between tasks, construct the conflict value weight matrix, establish undirected edge connection relationship according to the conflict value between nodes, label the conflict edge weight, insert the conflict edge into the initial task graph node set, establish the topological connection structure between nodes, and obtain the task node topological structure.

[0066] Based on the node conflict intensity distribution values, the system analyzes the conflict connections between tasks. Specifically, the system creates a 50x50 zero matrix as the initial form of the conflict value weight matrix, with the row and column indices corresponding to task numbers T01 to T50. Next, the system iterates through the set of node conflict intensity distribution values. For each task pair and its conflict intensity value, for example, the conflict intensity value of (T03, T08) is 0.2096. The system fills in the value 0.2096 at the 3rd row, 8th column and the 8th row, 3rd column positions of the matrix. After filling in the data for all 112 pairs of conflicting tasks, this matrix becomes the conflict value weight matrix. Subsequently, the system uses this weight... The system constructs a graph structure using matrices, treating the 50 tasks as 50 independent nodes. It iterates through the elements in the weight matrix that have a value greater than 0. If the value at position (i,j) is w (w>0), an undirected edge is created between task nodes Ti and Tj, and the weight of the edge is assigned to w. For example, an undirected edge with a weight of 0.2096 is created between T03 and T08. The system adds all such conflicting edges to the original initial task graph, which only contains nodes and sequential dependencies (if any). Finally, a set of nodes or simple dependency lines is transformed into a complex network rich in conflict information. This network is the task node topology.

[0067] Please see Figure 3 The specific steps for obtaining priority data for scheduling nodes are as follows:

[0068] S211: Obtain the task node topology, collect the stage sequence number and the number of upstream dependent nodes for each task node, determine the path depth of each node in the task dependency chain, and generate path depth distribution values.

[0069] To obtain the task node topology, the system first collects the "stage sequence number" attribute for each task node in the structure. This number is preset during the project planning phase and represents the logical order of the task in the standard process flow. For example, the stage sequence number of T25 (roadbed excavation) is 3, and the number of T28 (base course paving) is 4. Next, the system traverses the topology and counts the number of times each node appears as the endpoint of an edge. This is the number of upstream dependent nodes. For example, if there is a dependent connection from T25 to T28, then the number of upstream dependent nodes of T28 is at least 1. The system uses this information to determine the number of upstream dependent nodes. To determine the path depth of each node in the task dependency chain, the specific execution process is as follows: Starting from the initial node with no upstream dependencies (depth 0), a breadth-first search or similar traversal method is used to move downstream nodes layer by layer. The path depth is incremented by 1 for each layer moved. If a node has multiple upstream dependencies, its path depth is the maximum of the path depths of all upstream nodes plus 1. For example, if T01 has a depth of 0, T02 depends on T01, then T02 has a depth of 1; if T03 depends on T01, then T03 has a depth of 1; and if T04 depends on both T02 and T03, then T04 has a path depth of 1. The system performs this calculation on all 50 nodes to obtain the path depth value for each node, for example, {T01:0,T02:1,T03:1,T04:2,…,T50:8}. This set is the path depth distribution value.

[0070] S212: Based on the path depth distribution value, detect the set of nodes in the region to which the task space belongs, compare the number of nodes and the area of ​​space in each region, calculate the node density of the region, and associate the density data with the path depth distribution value to obtain the node density feature parameters;

[0071] Based on the path depth distribution values, the system divides the geospatial area of ​​the construction project into a grid. For example, the entire 2 square kilometer construction area is divided into 10 rectangular areas of 200 meters x 1000 meters (numbered R1 to R10). Then, the system detects the geographic coordinates of each task node and assigns it to the corresponding spatial area, forming a node set. For example, the node set contained in area R3 is {T04, T05, T11, T12, T15}. The system compares the number of nodes in each area with the area of ​​that area to calculate the node density of each area. Specifically, the calculation is the number of nodes in the area divided by the area. For area R3, the number of nodes is 5, and the area is 0.2 square kilometers, therefore its node density is... The system performs this calculation on all 10 regions per square kilometer, obtaining a set of density data {R1:15,R2:20,R3:25,…}. Finally, the system associates this density data with the path depth distribution value. Specifically, it adds a "regional density" attribute to each task node and assigns the density value of the region where the node is located to the target node. For example, if T04 is located in region R3, then the regional density attribute value of T04 is 25. In this way, the density feature parameters of all nodes are obtained.

[0072] S213: Call the node density feature parameters, calculate the task node priority ranking score based on the path depth distribution value of each task node, the number of upstream dependencies, and the density of the node's region, adjust the connection relationship of task nodes in the topology, reconstruct the node connection structure in the task graph, and generate scheduling node priority data.

[0073] The system calls the node density feature parameters and calculates the task node priority ranking score based on three indicators: the path depth distribution value of each task node, the number of upstream dependencies, and the density of the node's region. The score calculation process is a weighted sum of the three indicators, and the weights are set based on historical project data analysis, with path depth weighting being one of the indicators. Set to 0.5, weight of the number of upstream dependencies. Set to 0.3, regional density weight The weights were set to 0.2. These weights were determined through regression analysis of 10 completed projects. The analysis showed that path depth had the most critical impact on project delays (correlation coefficient 0.68), followed by dependencies (correlation coefficient 0.45), while regional density had a relatively smaller impact (correlation coefficient 0.31). The weight values ​​were obtained by normalizing the correlation coefficients. For example, for task node T04, its path depth is 2, the number of upstream dependencies is 2, and the regional density is 25. Before calculation, each indicator needs to be normalized. Assuming that the maximum path depth among all current tasks is 8, the maximum number of dependencies is 5, and the maximum density is 30, then the normalized value for T04 is: Depth Dependence on quantity density Then the priority ranking score for T04 is The system performs the same scoring calculation on all 50 nodes to obtain a complete score list. Subsequently, the system sorts all task nodes from high to low according to this score and adjusts the connection relationship of task nodes in the topology based on this new order. Specifically, it rearranges the node positions at the visualization or data structure level so that nodes with higher scores are logically placed in a higher priority position, thereby reconstructing the node connection structure in the task graph and generating scheduling node priority data.

[0074] Please see Figure 4 The specific steps for obtaining the impact indicators of path delay are as follows:

[0075] S311: Call the priority data of the scheduling node, extract the nodes in each task path, collect the duration of each node, personnel attendance stability, equipment scheduling waiting time and environmental interference conditions, analyze the actual delay impact of each node in this path, calculate the delay contribution ratio of the node, and generate the node delay contribution coefficient.

[0076] The system first extracts all key task paths from the reconstructed task graph, such as path P1: T01->T04->T11->T25->T48, by calling the priority data of scheduling nodes. Then, it collects detailed attributes for each node in path P1. Taking node T11 as an example, its planned duration is 43200 seconds, the personnel attendance stability coefficient is 0.95 (obtained by querying the associated personnel management system and calculating the attendance rate of the job type required for this task over the past month), the predicted equipment scheduling waiting time is 3600 seconds (based on the average waiting time of historical similar equipment scheduling records), and the environmental interference condition coefficient is 0.8 (quantified based on weather forecasts and surrounding traffic control information). (1 represents no interference, 0 represents complete interruption, and 0.8 indicates an expected 20% efficiency loss). The system analyzes the actual delay impact of T11 on this path. The calculation method is to comprehensively estimate a total delay risk duration based on various factors. For example, delay risk duration = (planned duration / (personnel stability * environmental interference coefficient)) + equipment scheduling waiting time - planned duration = (43200 / (0.95 * 0.8)) + 3600 - 43200 = 56842 + 3600 - 43200 = 17242 seconds. Next, the system calculates the total delay risk duration of all nodes on path P1. Assuming the total is 54000 seconds, the delay contribution ratio of node T11 is... The system repeats this calculation for all nodes in path P1 and all nodes in other task paths to generate the delay contribution coefficient for all nodes.

[0077] S312: Based on the node delay contribution coefficient, filter intersecting nodes by judging the number of positions of the same node in multiple paths, and statistically analyze the path distribution characteristics of intersecting nodes to obtain the distribution parameters of intersecting nodes;

[0078] Based on the node delay contribution coefficient, the system traverses all identified task paths and filters intersecting nodes by determining the occurrence of the same task node in different paths. Specifically, the system creates a hash table with task node IDs as keys and the list of appearing paths as values. When traversing path P1={T01,T04,T11,T25}, the system records {T01:[P1],T04:[P1],T11:[P1],T25:[P1]} in the hash table. When traversing another path P2={T02,T04,T15,T25}, the hash table is updated to {T01:[P1]}. After traversing the path list, the system selects nodes whose path list length is greater than 1. These are the intersection nodes. In this example, T04 and T25 are identified as intersection nodes. The system then counts the path distribution characteristics of these intersection nodes. For example, for T04, the system records that it appears in 2 paths and is the 2nd position in both paths. The path count of T04 is 2, and the position distribution is {P1:2,P2:2}. This is the intersection node distribution parameter.

[0079] S313: Call the intersection node distribution parameters, determine the delay risk distribution structure of each intersection node under multiple paths, classify and identify the risk distribution density, output the risk level index of each intersection node, and generate the path delay impact index.

[0080] By calling the distribution parameters of the convergence nodes, the system determines the delay risk distribution structure of each convergence node across multiple paths. Taking convergence node T25 as an example, it exists in both paths P1 and P2. The system retrieves the delay contribution coefficient of T25 in P1 (e.g., 0.25) and the delay contribution coefficient in P2 (e.g., 0.35). The system categorizes and labels these risk distribution densities. The labeling rules are preset as follows: if the average delay contribution coefficient of a node across multiple paths exceeds 0.3, or the contribution coefficient in any single path exceeds 0.4, its risk distribution density is labeled as "high"; if the average is between 0.15 and 0.3, it is labeled as "medium"; and if it is below 0.15, it is labeled as "low". The average coefficient of T25 is... The system meets the "medium" risk standard. At the same time, the maximum contribution coefficient of T25 is 0.35, which does not exceed 0.4. Therefore, its risk distribution density is initially identified as "medium". Based on this, the system outputs the risk level index of each intersection node, mapping "high" to risk level 3, "medium" to 2, and "low" to 1. Therefore, the risk level index of T25 is 2. The system completes this operation for all identified intersection nodes and finally generates a list containing each intersection node and its corresponding risk level. This is the path delay impact index.

[0081] Please see Figure 5 The specific steps for obtaining scheduling adaptation level data are as follows:

[0082] S411: Based on the path delay impact index, analyze the task process type in each affected path, extract the equipment number associated with the corresponding task, compare the matching frequency of the equipment participating in the process with the process category required by the current task, and obtain process consistency label data.

[0083] Based on the path delay impact index, the system identifies paths containing nodes with risk levels of 3 and 2 as heavily affected paths, such as paths P1 and P2 containing task T25. The system further analyzes the task types within path P1. Assuming task T25 in P1 is the "asphalt surface paving" task, the system extracts the equipment numbers associated with T25, which are {D-007 (paver), D-015 (roller), D-016 (roller)}. Subsequently, the system queries the equipment historical database to compare the participation of these three machines in the "asphalt surface paving" task. According to database records, D-007 participated 35 times, D-015 participated 50 times, and D-016 participated 12 times. At the same time, the system checks the consistency of the process category required by the current task T25 and confirms that "asphalt pavement laying" is a single and clear process category. The system associates the historical matching frequency of each piece of equipment with the process category (which is consistent here) to form process matching data {D-007:35, D-015:50, D-016:12}, and attaches a consistency tag to obtain the process consistency tag data.

[0084] S412: Call the process consistency tag data to determine the deviation of the equipment's response time in the scheduling start-up phase and the operation error in the execution phase, filter the equipment numbers with abnormal response time and operation error deviation, and establish a response deviation abnormality coefficient table;

[0085] The system retrieves process consistency tag data and, for equipment D-007, D-015, and D-016, accesses their performance monitoring logs to determine the equipment's response time during the scheduling initiation phase and the operational error deviation during the execution phase. Response time is defined as the time difference between the issuance of the scheduling command and the equipment's first movement recorded by GPS. Operational error deviation is defined as the percentage deviation of the average laying thickness measured by the laser scanner from the design value of 8 cm. The system sets an abnormal threshold of 900 seconds for response time and ±5% for operational error deviation. These thresholds are based on statistical analysis of historical performance data for this equipment model, taking the 95th percentile of the response time distribution (880 seconds, rounded to 900 seconds) and the 95th percentile of the error distribution. (4.8%, rounded to 5%) is used as the boundary for judging anomalies. The system detected that the average response time of D-007 is 450 seconds and the operation error is +1.2%; the average response time of D-015 is 600 seconds and the operation error is -0.8%; the average response time of D-016 is 1100 seconds and the operation error is +6.5%. Since the response time of D-016 (1100 seconds) is greater than 900 seconds and the operation error (6.5%) is greater than 5%, the system filters out D-016 as the device number with abnormal response time and operation error deviation and records it in a response deviation anomaly coefficient table. The table records the device number and the corresponding anomaly coefficient. The coefficient is defined as the proportion exceeding the threshold. The response anomaly coefficient of D-016 is... The error anomaly coefficient is .

[0086] S413: Based on the response offset anomaly coefficient table, extract the process matching frequency, response time, and error offset rate for each device, using the following formula:

[0087] ;

[0088] Calculate the scheduling adaptability score for each device and generate scheduling adaptability score data;

[0089] in, The scheduling adaptability rating represents the equipment's ability to adapt to the current task conditions, and is used to measure the degree of matching between the equipment and the current task. The equipment number is The normalized value of the process matching frequency represents the ratio of the frequency of its participation in the current task process to the maximum frequency among all devices. It is obtained by calculating the ratio of the number of times the device appears in this process type to the maximum number of times it appears. The equipment number is The normalized response time value represents the ratio of the device's startup response time to its maximum response time. It is calculated by collecting the actual time from the issuance of the scheduling command to the start of the action and comparing it with the longest response time among all devices. The equipment number is The operation error offset rate represents the percentage of error offset by the equipment during task execution. It is calculated by recording the ratio of the operation error generated by the equipment during task execution to the allowable error threshold. This represents the total number of devices required for the current task to be matched, indicating the total number of devices involved in the scoring calculation. For device index variables, it represents the number of devices participating in the scheduling evaluation. One device;

[0090] Based on the response offset anomaly coefficient table, the system excludes abnormal equipment D-016. For the remaining equipment D-007 and D-015, the system extracts their process matching frequency, response time, and error offset rate, and then uses the formula... Calculate the scheduling adaptability level score;

[0091] Table 1. Raw Data of Equipment Performance

[0092] ;

[0093] Table 1 lists the raw performance data of the equipment involved in the scoring calculation. The parameters and calculation process in the formula are explained below. This scheduling adaptability level score... By employing a structure combining multiplication and addition, the multi-dimensional performance of the equipment and the frequency of process matching are comprehensively evaluated. As a basic weight, it reflects the advantages of the equipment's historical experience, and the normalized value of the response time. The reciprocal of this factor amplifies the disadvantage of slow response time, while also penalizing the operational error offset rate. The processing, through taking the absolute value and square root, aims to smooth out the impact of errors while ensuring that both positive and negative errors are treated as negative factors, with larger errors resulting in more significant penalties. The final summation and averaging (divided by) This integrates the evaluation results from multiple devices into an overall adaptability metric for the current task. For device index variables, This represents the total number of devices participating in the scoring, which is 2 here. For device D-007 ( ) and D-015 ( ): The normalized value for equipment process matching frequency is obtained by calculating the ratio of the equipment matching frequency to the maximum frequency (50 times) among all equipment. , , This is the normalized value of the device response time, calculated as the ratio of the device response time to the maximum response time (600 seconds). , , For the job error offset rate, use the decimal value after percentage conversion. , Substitute the numerical values ​​into the formula to perform the calculation:

[0094] The individual score for D-007 is as follows: ;

[0095] The individual score for D-015 is as follows: ;

[0096] The final scheduling fitness level score is: ;

[0097] The scheduling adaptability rating is a numerical measure of the adaptability of multiple construction devices within a specific scheduling path or task. This rating comprehensively considers the device's historical performance in task adaptation, scheduling response speed, and error control capabilities during operation. By evaluating these core factors affecting scheduling quality in a coordinated manner, a scheduling adaptability index is derived, which can be used for sorting, filtering, and setting resource priority. Specifically, the rating result can be directly used to construct task-device adaptability maps, resource priority calling tables, or automatic scheduling sorting tables. A higher rating indicates stronger availability of the device under the current task conditions, making it more suitable for priority deployment. Driving the scheduling execution process with this parameter can effectively reduce equipment idling and waiting, lower scheduling failure rates, and improve construction progress stability and task execution efficiency. The results indicate that the scheduling adaptability score of the device group consisting of devices D-007 and D-015 for task T25 is 0.96585. The system's preset adaptation level range is: [0.9, 1.0] is highly adapted, [0.7, 0.9) is moderately adapted, and below 0.7 is low adapted. The score of 0.96585 is in the highly adapted range, which means that this device group is very suitable for performing the current task. The system associates this score of 0.96585 with task T25 to form scheduling adaptability level data. The scheduling adaptability rating is a numerical measure of the adaptability of multiple construction devices in a specific scheduling path or task. This rating comprehensively considers the device's historical performance in process adaptation, scheduling response speed, and error control capabilities during operation. By evaluating these core factors that affect scheduling quality in a coordinated manner, a scheduling adaptability index is derived that can be used for sorting, filtering, and setting resource priority. In practice, the rating result can be directly used to construct task-device adaptability maps, resource priority calling tables, or automatic scheduling sorting tables. The higher the rating, the stronger the availability of the device under the current task conditions, and the more suitable it is for priority use. By driving the scheduling execution process with this parameter, it is possible to effectively reduce equipment idling and waiting, reduce scheduling failure rate, and improve the stability of construction progress and task execution efficiency.

[0098] Please see Figure 6 The specific steps for obtaining the resource scheduling and arrangement results are as follows:

[0099] S511: Based on the scheduling adaptation level data, analyze the participation frequency of the currently assigned tasks for each device, detect the duration of unassigned tasks for each device, determine the scheduling interval period between consecutive tasks of the device, and generate device scheduling interval parameters.

[0100] Based on the scheduling adaptation level data, the system analyzes the participation frequency of currently assigned tasks and the duration of unassigned tasks for devices D-007 and D-015. The system queries the scheduling table and finds that D-007 has been assigned 2 tasks in the next 24 hours, totaling 10 hours, while D-015 has been assigned 1 task, totaling 6 hours. The system detects a 4-hour idle window between the two assigned tasks for D-007, which represents the duration of its unassigned tasks. D-015 has 18 hours of idle time after completing its current task. The system determines the scheduling interval period between consecutive tasks, i.e., the idle window duration. The scheduling interval period for D-007 is 4 hours, and for D-015 it is 18 hours. The system integrates this information to generate device scheduling interval parameters, such as {D-007:4h, D-015:18h}.

[0101] S512: Call the device scheduling interval parameter, compare the matching degree between the device's calling order position and the scheduling adaptation level in the current scheduling cycle, determine the consistency between the order position and the adaptation level, adjust the order of the devices in the scheduling order table, and generate the device order adjustment sequence.

[0102] The system calls the equipment scheduling interval parameter and compares the matching degree between the equipment's scheduling order position and scheduling adaptability level within the current scheduling cycle. The system queries the current pre-scheduled order table and finds that for the next new task T26 (similar to T25), D-007 is ranked 5th and D-015 is ranked 2nd. However, according to the calculation in S413, D-015's single-item adaptability (1.0040) is higher than D-007's (0.9277), indicating a discrepancy between the order position and adaptability level. Specifically, D-015, with higher adaptability, is not ranked first. The system determines this to be inconsistent and requires further action. The adjustment logic is as follows: Among candidate devices with scheduling intervals (all greater than 0), priority is given to improving the ranking of devices with higher single-item adaptability scores. Since D-015 has the highest adaptability score and a sufficient scheduling interval of 18 hours, the system promotes its ranking to 1st. D-007 has the next highest adaptability score and a 4-hour interval, so the system promotes its ranking to 2nd. The devices in the original ranking list that are 1st and 2nd are then moved up. The system generates a sequence containing specific adjustment instructions, namely {Instruction 1: Move D-015 to ranking 1; Instruction 2: Move D-007 to ranking 2}. This is the device ranking adjustment sequence.

[0103] S513: Adjust the sequence according to the equipment order, update the arrangement structure of multiple equipment resources in the scheduling instruction table, integrate the order number of each equipment with the instruction mapping, and obtain the resource scheduling arrangement result;

[0104] Based on the device sorting and adjustment sequence, the system executes instructions to update the arrangement structure of multiple device resources in the scheduling instruction table. The original device call order for task T26 in the scheduling instruction table was [D-010, D-015, D-021, D-022, D-007, ...]. After the system executes the adjustment sequence, the new arrangement structure becomes [D-015, D-007, D-010, D-021, D-022, ...]. The system further integrates the new sorting number of each device with specific scheduling instructions. For example, a new instruction is generated for D-015: "Task T26, sorting 1, estimated start time 1765603200", and for D-007, the instruction is generated: "Task T26, sorting 2, estimated start time 1765603200". By completing this operation on all affected devices, an optimized, detailed, and directly executable instruction list is finally obtained; this is the resource scheduling arrangement result.

[0105] An AI-based road construction resource optimization and allocation system is used to execute the aforementioned AI-based road construction resource optimization and allocation method. The system includes:

[0106] The node conflict identification module calls up road construction task information, analyzes the work area boundary and work period of task nodes, determines the spatial overlap and temporal concurrency between adjacent nodes, compares the resource codes of task equipment configuration items, determines the conflict intensity of task nodes, and generates the task node topology.

[0107] The task priority ranking module analyzes the task stage order, upstream dependencies and spatial density based on the task node topology, calculates the task node priority ranking score, rearranges the node order according to the task number, reconstructs the connection structure of the task graph, and generates scheduling node priority data.

[0108] The path delay assessment module calls the priority data of the scheduling nodes, extracts the duration, attendance stability, equipment waiting time and interference conditions of the path nodes, calculates the delay contribution ratio, judges the intersection nodes and the distribution of delay risks, and generates path delay impact indicators.

[0109] The scheduling level determination module analyzes the correspondence between the equipment number and the task process type involved in the affected path based on the path delay impact index, determines the equipment response time and execution error deviation, filters abnormal equipment, calculates the scheduling adaptability level score, and generates scheduling adaptability level data.

[0110] The resource order adjustment module analyzes the device task frequency and idle time based on the scheduling adaptation level data, judges the scheduling interval, compares the matching degree between the call order and the adaptation level, adjusts the device scheduling order, updates the arrangement structure of resources in the scheduling instruction table, and generates the resource scheduling arrangement result.

[0111] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. An AI-based method for optimizing the allocation of road construction resources, characterized in that, Includes the following steps: S1: Call the road construction task information, analyze the work area boundary and work period of the task node, determine the spatial overlap and temporal concurrency between adjacent nodes, compare the resource codes of the task equipment configuration items, determine the conflict intensity of the task node, and generate the task node topology. The specific steps for obtaining the task node topology are as follows: S111: Obtain road construction task information, analyze the work area boundary information and work time plan of task nodes, determine the area overlap and work time coverage between adjacent nodes, and obtain task overlap combination. S112: Invoke the task overlap combination, extract the equipment configuration information table of each node in the combination, compare the resource codes in the task equipment configuration items, identify resource configuration duplication, and use the following formula based on the area of ​​regional overlap, the proportion of overlapping operation time, and the proportion of resource code duplication: ； Calculate the conflict intensity value of the task combination, map the conflict intensity value to the task node combination, and generate the node conflict intensity distribution value; in, This represents the conflict intensity value of the task combination. This is the normalized value of the area of ​​intersection of regions. This is the normalized value of the overlap ratio of task times. Configure the repetition ratio for task combination resources. This is the normalized value of the total number of tasks in the current task set. The normalized task distribution density per unit area; S113: Based on the node conflict intensity distribution value, analyze the conflict connection relationship between tasks, construct a conflict value weight matrix, establish undirected edge connection relationship according to the conflict value between nodes, label the conflict edge weight, insert the conflict edge into the initial task graph node set, establish the topological connection structure between nodes, and obtain the task node topological structure. S2: Based on the task node topology, analyze the task phase sequence, upstream dependencies and spatial density, calculate the task node priority ranking score, rearrange the node order according to the task number, reconstruct the connection structure of the task graph, and generate scheduling node priority data. S3: Call the priority data of the scheduling node, extract the duration, attendance stability, equipment waiting time and interference conditions of the path node, calculate the delay contribution ratio, determine the intersection node and delay risk distribution, and generate path delay impact index; S4: Based on the path delay impact index, analyze the correspondence between the equipment number and the task process type involved in the affected path, determine the equipment response time and execution error deviation, screen abnormal equipment, calculate the scheduling adaptation level score, and generate scheduling adaptation level data.

2. The AI-based road construction resource optimization allocation method according to claim 1, characterized in that, The task node topology includes spatial intersection relationship type, time-overlapping task group number, and resource code repetitive mapping relationship. The scheduling node priority data specifically includes stage sequence number, path depth identifier, and spatial distribution density. The path delay impact index includes delay contribution ratio distribution, node intersection location index, and delay risk associated path number. The scheduling adaptation level data specifically includes equipment task matching frequency, scheduling response time label, and operation error offset amplitude.

3. The AI-based road construction resource optimization allocation method according to claim 1, characterized in that, The specific steps for obtaining the priority data of the scheduling node are as follows: S211: Obtain the topology of the task nodes, collect the stage sequence number and the number of upstream dependent nodes of each task node, determine the path depth of each node in the task dependency chain, and generate path depth distribution values. S212: Based on the path depth distribution value, detect the node set of the region to which the task space belongs, compare the number of nodes and the space area of ​​each region, calculate the node density of the region, and associate the density data with the path depth distribution value to obtain the node density feature parameters; S213: Call the node density feature parameters, calculate the task node priority ranking score based on the path depth distribution value, the number of upstream dependencies, and the density of the area where the node is located for each task node, adjust the connection relationship of the task nodes in the topology, reconstruct the node connection structure in the task graph, and generate scheduling node priority data.

4. The AI-based road construction resource optimization allocation method according to claim 3, characterized in that, The specific steps for obtaining the path delay impact index are as follows: S311: Call the priority data of the scheduling node, extract the nodes in each task path, collect the duration of each node, personnel attendance stability, equipment scheduling waiting time and environmental interference conditions, analyze the actual delay impact of each node in this path, calculate the delay contribution ratio of the node, and generate the node delay contribution coefficient. S312: Based on the node delay contribution coefficient, by judging the number of positions of the same node in multiple paths, intersecting nodes are filtered, and the path distribution characteristics of intersecting nodes are statistically analyzed to obtain the distribution parameters of intersecting nodes; S313: Call the intersection node distribution parameters, determine the delay risk distribution structure of each intersection node under multiple paths, classify and identify the risk distribution density, output the risk level index of each intersection node, and generate the path delay impact index.

5. The AI-based road construction resource optimization allocation method according to claim 4, characterized in that, The specific steps for obtaining the scheduling adaptation level data are as follows: S411: Based on the path delay impact index, analyze the task process type in each affected path, extract the equipment number associated with the corresponding task, compare the matching frequency of the equipment participating in the process with the process category required by the current task, and obtain process consistency label data. S412: Call the process consistency tag data to determine the deviation of the equipment's response time in the scheduling start-up phase and the operation error in the execution phase, filter the equipment numbers with abnormal response time and operation error deviation, and establish a response deviation abnormality coefficient table; S413: Based on the response offset anomaly coefficient table, extract the process matching frequency, response time and error offset rate of each device, calculate the scheduling adaptation level score of each device, and generate scheduling adaptation level data.

6. The AI-based road construction resource optimization allocation method according to claim 1, characterized in that, The method further includes: S5: Based on the scheduling adaptation level data, analyze the device task frequency and idle time, determine the scheduling interval, compare the matching degree between the call order and the adaptation level, adjust the device scheduling order, update the arrangement structure of resources in the scheduling instruction table, and generate resource scheduling arrangement results. The resource scheduling arrangement result includes the resource number sorting position, the update order in the instruction table, and the device scheduling priority identifier.

7. The AI-based road construction resource optimization allocation method according to claim 6, characterized in that, The specific steps for obtaining the resource scheduling and arrangement results are as follows: S511: Based on the scheduling adaptation level data, analyze the participation frequency of the currently assigned tasks for each device, detect the duration of unassigned tasks for each device, determine the scheduling interval period between consecutive tasks of the device, and generate device scheduling interval parameters. S512: Call the device scheduling interval parameter, compare the matching degree between the device's scheduling position and the scheduling adaptation level in the current scheduling cycle, determine the consistency between the scheduling position and the adaptation level, adjust the arrangement order of the devices in the scheduling sorting table, and generate a device sorting adjustment sequence. S513: Adjust the sequence according to the device sorting, update the arrangement structure of multiple device resources in the scheduling instruction table, integrate the sorting number of each device with the instruction mapping, and obtain the resource scheduling arrangement result.

8. An AI-based road construction resource optimization and allocation system, characterized in that, The system is used to implement the AI-based road construction resource optimization allocation method according to any one of claims 1-7, and the system includes: The node conflict identification module calls up road construction task information, analyzes the work area boundary and work period of task nodes, determines the spatial overlap and temporal concurrency between adjacent nodes, compares the resource codes of task equipment configuration items, determines the conflict intensity of task nodes, and generates the task node topology. The task priority ranking module analyzes the task stage order, upstream dependencies and spatial density based on the task node topology, calculates the task node priority ranking score, rearranges the node order according to the task number, reconstructs the connection structure of the task graph, and generates scheduling node priority data. The path delay assessment module calls the priority data of the scheduling nodes, extracts the duration, attendance stability, equipment waiting time and interference conditions of the path nodes, calculates the delay contribution ratio, judges the intersection nodes and the distribution of delay risks, and generates path delay impact indicators. The scheduling level determination module analyzes the correspondence between the equipment number and the task process type involved in the affected path based on the path delay impact index, determines the equipment response time and execution error deviation, filters abnormal equipment, calculates the scheduling adaptability level score, and generates scheduling adaptability level data. The resource order adjustment module analyzes the device task frequency and idle time based on the scheduling adaptation level data, judges the scheduling interval, compares the matching degree between the call order and the adaptation level, adjusts the device scheduling order, updates the arrangement structure of resources in the scheduling instruction table, and generates the resource scheduling arrangement result.

Images