Enterprise resource scheduling optimization method based on artificial intelligence and large model algorithm

Abstract

The invention provides an enterprise resource scheduling optimization method based on artificial intelligence and a large model algorithm, and the method comprises the steps: recalculating the earliest start time and the latest completion time of each operation node in an affected downstream project according to a project dependency graph and a resource constraint condition, and obtaining an optimized time window distribution scheme; combining the personnel candidate set and the time window distribution scheme to calculate a matching degree score of the personnel skill level and the engineering demand characteristics, and if the matching degree score is higher than a set standard value, generating a personnel allocation suggestion list; and according to the target personnel flow scheme, updating the personnel configuration table and the skill structure distribution of each project, recalculating a project progress predicted value and a resource utilization rate index, generating a dynamic personnel redistribution execution scheme, and updating system configuration parameters.

Description

Technical Field

[0001] The present invention relates to the field of information technology, and in particular to an enterprise resource scheduling optimization method based on artificial intelligence and a large model algorithm. Background Art

[0002] Power grid projects are core infrastructure for modern energy supply, and their construction and maintenance are directly linked to economic development and livelihood security. Project delays, a common problem in power grid construction, not only impact project progress but can also lead to resource waste and increased costs. Current strategies for addressing project delays often focus on static resource scheduling or single-factor optimization, such as adjusting construction periods or increasing staffing. However, these approaches struggle to adapt to the multi-dimensional changes caused by delays, particularly regarding workforce skill matching and the dynamic adjustment of work time windows. Existing solutions often overlook the supply and demand balance between personnel of varying skill levels in complex projects, resulting in poor resource utilization and delayed critical tasks. The core challenge of project delays lies in the dynamic imbalance in the workforce's skill mix. When plans are disrupted by external factors, a direct consequence is that some highly skilled personnel may be idle due to adjustments to their original assignments. At the same time, equally critical tasks requiring lesser skills may face delays due to staffing shortages. This sudden imbalance between skilled resources and current task demands not only directly reduces overall operational efficiency but, more critically, significantly exacerbates the difficulty of reallocating time windows required by the disruption. Specifically, schedule disruptions often mean that existing task execution time windows must be adjusted, and tasks need to be reordered under new time constraints. However, skills imbalances prevent available human resources from flexibly and efficiently adapting to these newly adjusted task requirements and time windows. This results in difficulties finding suitably skilled personnel to execute even when new time windows are available, dramatically increasing the complexity of scheduling. Therefore, designing an intelligent strategy that dynamically adjusts skill structure and time window allocation based on personnel capabilities and project requirements has become a key issue in optimizing resources and ensuring schedule assurance during power grid project delays. Summary of the Invention

[0003] The present invention provides an enterprise resource scheduling optimization method based on artificial intelligence and a large model algorithm, which mainly includes:

[0004] Obtain the current progress and delay information of each power grid project, analyze the completion status of project nodes and the remaining workload, and calculate the number of days of delay for each project and the number of affected downstream projects, power supply areas, and users based on standard construction period parameters to obtain a project dependency diagram;

[0005] Based on the number of days of delay and the scope of impact, the quantity demanded and skill level requirements for different types of work at each project node are calculated. The relationship between project task types and skill levels is analyzed and matched to obtain a supply-demand difference matrix for the number of high-skilled personnel shortages and the number of low-skilled personnel redundancies.

[0006] Analyze the supply-demand discrepancy matrix and identify gaps. Compare the supply and demand values ​​of each skill level to identify gaps. If the shortage of highly skilled personnel exceeds the threshold, query the personnel skills database to determine the candidate set and the idle highly skilled personnel resource pool.

[0007] Based on the project dependency graph and resource constraints, the earliest start time and latest completion time of each job node in the affected downstream projects are recalculated to obtain an optimized time window distribution plan.

[0008] Combine the candidate set and the time window distribution scheme to calculate the matching score between the personnel skill level and the engineering requirements. If the matching score is higher than the set standard value, a list of recommended personnel deployment is generated.

[0009] Calculate the shortest deployment path and time cost for cross-project personnel mobility based on the recommended staff deployment list. Determine the target personnel mobility plan based on the transportation distance between project sites, deployment costs, skill matching, and the pool of available highly skilled personnel.

[0010] Update the staffing table and skill structure distribution of each project according to the target personnel flow plan, recalculate the project progress forecast and resource utilization indicators, generate a dynamic personnel reallocation execution plan and update the system configuration parameters.

[0011] The technical solution provided by the embodiment of the present invention may have the following beneficial effects:

[0012] The present invention discloses an enterprise resource scheduling optimization method based on artificial intelligence and large model algorithms. By acquiring project progress and extension information, analyzing node completion status and remaining workload, calculating the number of extension days and the scope of impact, constructing a project dependency graph, and identifying the gap in high-skilled personnel and the redundancy of low-skilled personnel based on the skill level correspondence table, a supply and demand difference matrix is ​​generated. In response to the gap in high-skilled personnel, the present invention activates a cross-project personnel flow mechanism, adopts a time window reconstruction algorithm to optimize the time distribution of job nodes, combines the personnel skills with the project demand matching score, and generates a deployment recommendation list. Further considering factors such as the distance between projects and deployment costs, the optimal personnel flow plan is determined, the project personnel configuration and skill structure are updated, and the project progress and resource utilization are re-forecasted. The present invention realizes the precise dynamic deployment of power grid engineering personnel, improves the efficiency of human resource utilization and the level of project progress management. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1This is a flow chart of an enterprise resource scheduling optimization method based on artificial intelligence and large model algorithms of the present invention.

[0014] Figure 2 This is a schematic diagram of an enterprise resource scheduling optimization method based on artificial intelligence and large model algorithms according to the present invention. DETAILED DESCRIPTION

[0015] The technical solutions of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0016] like Figure 1-2 In this embodiment, an enterprise resource scheduling optimization method based on artificial intelligence and a large model algorithm may specifically include:

[0017] S101. Obtain the current progress and delay information of each power grid project, analyze the completion status of project nodes and the remaining workload, and calculate the number of days of delay for each project and the number of affected downstream projects, power supply areas, and users based on standard construction period parameters to obtain a project dependency diagram.

[0018] Obtain the planned start time, planned end time, and actual completion time of each power grid project, read the completion status identifier of each node in the project node list, calculate the ratio of the number of completed nodes to the total number of nodes to obtain the progress rate, query the standard working hours corresponding to each unfinished node from the project quota library, and accumulate the standard working hours in the order of nodes to obtain the remaining work. For each project, calculate the difference between the actual completion time and the planned end time. If the difference is greater than the preset extension judgment threshold, the difference is determined to be the number of extension days. Query the predecessor and successor relationship data between projects based on the project number, extract all project numbers with the current project as the prerequisite as the downstream project list, and count the number of projects in the list to obtain the number of affected downstream projects. According to the project number in the downstream project list, query the power supply facility code corresponding to each downstream project, determine the affected power supply area code set through the mapping relationship between the power supply facility and the power supply area, query the total number of users in the corresponding area in the user archive database for each power supply area code, and accumulate the total number of users in each area to obtain the number of affected users. A directed graph structure is constructed based on the predecessor and successor relationships between engineering projects, in which nodes represent engineering projects and are marked with the number of delay days, the number of affected downstream projects, the power supply area code set and the number of users. Edges represent dependency relationships. The directed graph is traversed using a depth-first traversal algorithm to obtain an engineering dependency graph that includes the delay information and impact range of each project.

[0019] Specifically, in power grid project management, progress monitoring and delay analysis are key steps in ensuring stable power supply. By obtaining project timeline information, we can accurately understand the actual implementation status of each project.

[0020] Specifically, the planned start time is usually determined at project initiation, while the planned end time is determined based on the project's scale and complexity. The actual completion time requires real-time feedback from the construction site. The project milestone list includes key steps such as foundation construction, equipment installation, and line erection. Each milestone is clearly marked with a completion status, such as "Completed," "In Progress," or "Not Started." The progress rate is calculated as the number of completed milestones divided by the total number of milestones, providing a direct indicator of the overall project progress. Calculating the remaining work is more complex, requiring a query of the standard working hours corresponding to each uncompleted milestone from the engineering quota database. The engineering quota database is a long-standing data resource within the power industry, recording the standard working hours required for different types of engineering tasks.

[0021] For example, the standard labor time for erecting a 1km 110kV transmission line is 120 hours, and the standard labor time for installing a transformer is 48 hours. By summing the standard labor time for all unfinished nodes, the remaining workload for the project can be calculated, providing a basis for subsequent resource allocation. Delays are determined using a difference calculation method. When the actual completion time exceeds the planned end time and the difference is greater than a preset threshold, a delay is considered. This threshold setting takes into account the normal fluctuation range of construction periods, typically 3-5 days. The predecessor and successor relationship data between projects records the dependencies between projects. For example, substation construction must wait until the related transmission lines are completed before power supply commissioning can begin. By querying this relationship data, all downstream projects affected by the current delayed project can be quickly identified.

[0022] In one possible implementation, the impact range analysis of downstream projects needs to be combined with the actual layout of the power supply network. Each engineering project corresponds to specific power supply facilities, such as substations, switch stations, or distribution lines. By mapping the power supply facilities and the power supply area, the affected geographical range can be accurately located. The power supply area is usually divided according to administrative divisions or power supply areas, and each area has a unique code. The user profile database records the user information in each power supply area in detail, including different types of users such as residential users and industrial and commercial users. By querying and accumulating the area codes, the total number of affected users can be accurately counted.

[0023] It's important to note that the project dependency graph uses a directed graph data structure, which clearly displays the complex relationships between projects. The nodes in the graph represent not only the project itself but also key information such as the number of days delayed, the number of affected downstream projects, the set of power supply region codes, and the number of users. Edges represent the dependencies between projects, and their directions indicate the order in which projects should be executed. A depth-first traversal algorithm begins with the delayed project and traces the dependencies layer by layer, identifying all potentially affected downstream projects. This visual representation enables managers to quickly grasp the chain reaction of delays, providing comprehensive decision-making support for countermeasures, effectively reducing the risk of widespread power outages and ensuring the safe and stable operation of the power grid.

[0024] S102. Based on the number of days of delay and the affected scope, calculate the quantity demand and skill level requirements of different types of work at each project node, analyze and match the relationship between project task types and skill levels, and obtain the supply and demand difference matrix of the number of high-skilled personnel gaps and the number of low-skilled personnel redundancies.

[0025] Based on the number of days of delay, query the construction period adjustment parameter table to obtain the corresponding construction period compression ratio. Multiply the original planned construction period by the construction period compression ratio to obtain the compressed construction period. Extract the task list of each project node and the type of work identification of each task from the affected range. Query the pre-established standard table of work type personnel configuration to obtain the standard number of personnel for each type of work required per unit task volume. Divide the original planned construction period by the compressed construction period to obtain the task intensity coefficient. Multiply the number of standard personnel by the task intensity coefficient to obtain the number of personnel required for each type of work under an expedited state. To determine the number of personnel required for each type of work under an expedited state, read the technical difficulty level of each task from the task list. Based on the technical difficulty level, query the minimum skill level required to perform the task in the type of work skill requirement comparison table. Sorting the skill levels from high to low, cumulatively calculate the number of personnel required for each skill level and construct a personnel demand matrix with types of work as rows and skill levels as columns. Obtain a list of currently available personnel from the human resources database, read each person's job category and skill level certificate number, query the corresponding skill level value in the skill certification database based on the certificate number, perform grouping statistics based on job category and skill level value, and construct a personnel supply matrix with job category as rows and skill level as columns. Subtract the corresponding elements of the personnel demand matrix from the personnel supply matrix to obtain a supply-demand difference matrix. Categorize the elements in the supply-demand difference matrix based on the positive and negative values ​​of each element. Elements with a skill level greater than a preset advanced threshold and a positive difference are identified as the number of high-skilled personnel gaps. Elements with a skill level less than a preset primary threshold and a negative difference are identified as the number of unskilled personnel surpluses by taking the absolute value. The number of high-skilled personnel gaps and the number of unskilled personnel surpluses are extracted to form the supply-demand difference matrix.

[0026] Specifically, the construction period adjustment parameter table is an important basic data in power grid project management. It is established based on historical project experience and records the construction period compression ratio corresponding to different delay days.

[0027] Specifically, when the delay is between 1 and 7 days, the compression ratio is typically set at 0.9, meaning that work that originally took 10 days must be completed within 9. When the delay exceeds 15 days, the compression ratio may reach 0.7, reflecting a more urgent need to speed up the project. This tiered approach takes into account the practicality of project execution and avoids quality issues caused by excessive compression. The calculation of the task intensity coefficient reflects the changing staffing requirements under expedited conditions.

[0028] For example, a transmission line project originally scheduled for 30 days needs to be completed in 21 days due to delays at an upstream substation. The task intensity coefficient is 30 divided by 21, which is approximately 1.43. This means that a task originally requiring 10 linemen would require 14 linemen to complete on time under expedited conditions. The standard personnel allocation table for each type of task details the standard allocation for each task. For example, the standard allocation for each kilometer of 110kV line is 8 linemen, 4 electricians, and 2 crane operators.

[0029] In one possible implementation, technical difficulty levels are determined based on the complexity and safety risks of the task. These levels are divided into five categories: Level 1 represents routine work, while Level 5 represents highly challenging work in special environments. A table of job skill requirements specifies the minimum skill requirements for tasks of varying difficulty levels. For example, Level 3 live-line work must be performed by individuals holding a senior worker certificate or higher, while Level 1 basic civil engineering work can be performed by a junior worker.

[0030] It should be noted that the use of a cumulative statistical method in constructing the personnel demand matrix is ​​of great significance. The accumulation from high-skill levels to low-skill levels is because high-skilled personnel can perform low-skilled tasks, but not vice versa.

[0031] For example, a project requires five senior electricians, eight intermediate electricians, and twelve junior electricians. The actual demand matrix shows a requirement of five senior electrician positions, 13 intermediate electrician positions (5 + 8), and 25 junior electrician positions (5 + 8 + 12). This approach ensures flexibility in staffing. The skill certification database stores the skill level information of all personnel, with each certificate number associated with a unique skill level. Skill levels are assigned on a standard scale of 1-10, with 1-3 representing junior, 4-6 representing intermediate, 7-9 representing senior, and 10 representing special. The preset threshold for senior is typically set at 7, and the threshold for junior is set at 3. This grading is based on industry skill standards and actual job requirements. The generation of the supply-demand discrepancy matrix accurately identifies human resource gaps. If the required number of personnel for a certain skill level in a certain job type is 15, but only 10 are available, the discrepancy is 5, and these five personnel represent the number of positions available. Conversely, if the demand for entry-level linemen is 20 but the actual number of available workers is 35, the remaining 15 workers can be used to fill other in-demand positions. This matrix representation allows managers to intuitively understand the supply and demand situation for each type of work and skill level, providing accurate data support for subsequent personnel deployment and training plan development, effectively solving the problem of human resource allocation during expedited project times.

[0032] Based on the correspondence table between engineering task types and skill levels, the actual number of senior skilled personnel on duty in each project is counted, compared with the urgent labor demand of delayed engineering nodes, the personnel gap value of high-skilled positions is identified, and the personnel redundancy of low-skilled positions is counted. A two-dimensional supply and demand difference matrix including the number of gaps, redundancy and urgency is constructed according to the type of work and skill level.

[0033] Based on the table of correspondences between engineering task types and skill levels, the minimum skill level requirements for each task type are queried. A list of currently employed personnel is extracted from the personnel attendance data. Based on the personnel number, their skill level certificates are queried in the skill certification database. Grouping and statistics are performed by work type and skill level to obtain a statistical table of the number of employed personnel for each work type and skill level. For this statistical table of the number of employed personnel for each work type and skill level, the number of delayed days and the original staffing plan for each node are extracted from the data of delayed project nodes. The construction period compression rate is calculated by dividing the number of delayed days by the original planned construction period. The original staffing number is divided by the construction period compression rate to obtain the number of personnel required for each work type and skill level in an emergency situation. This number is subtracted from the corresponding position in the on-the-job personnel statistics table to obtain a table of the difference between supply and demand for personnel. The positive and negative values ​​in the personnel supply and demand difference table are classified. If the difference is greater than zero, there is a staffing shortage for that type of work and skill level, and the number of shortages is recorded. If the difference is less than zero, there is a staffing surplus for that type of work and skill level, and the absolute value is taken as the number of surpluses. At the same time, an urgency value is assigned based on the number of days of delay, with the greater the delay, the higher the urgency value. Using the type of work as the row index and the skill level as the column index, the number of shortages, the number of surpluses, and the urgency value are used as the three components of the matrix elements to construct a two-dimensional supply and demand difference matrix containing the number of shortages, the number of surpluses, and the urgency. , D ij G represents the supply and demand difference vector corresponding to job type i and skill level j, ij Indicates the number of gaps at this position, R ij Indicates the amount of redundancy at this location, U ij This formula defines the three-component structure of each element in the two-dimensional supply-demand difference matrix.

[0034] Specifically, the preset correspondence table between engineering task types and skill levels is the core basic data for power grid engineering human resource management. It records in detail the minimum requirements for the skill levels of the executors of different types of tasks.

[0035] Specifically, the table divides engineering tasks into multiple categories, including civil construction, electrical installation, line installation, and equipment commissioning. Each category is further subdivided into specific tasks. Skill levels are assigned on a five-level scale, from entry-level worker to senior technician, with clear skill certification standards for each level.

[0036] For example, the installation of the main transformer in a 220kV substation requires at least senior workers or above, while ordinary cable laying tasks can be performed by junior workers.

[0037] In one possible implementation, counting the number of on-duty personnel requires integrating multiple data sources. Personnel attendance data records each employee's real-time attendance, including regular work, leave, and assignments. The skills certification database stores all personnel's skill level certificates, each with a unique number and expiration date. By linking these two data sources by personnel number, we can accurately calculate the number of personnel currently available at each skill level. This statistical approach avoids counting personnel on leave or assignment as available resources, improving the accuracy of manpower allocation. Calculating the schedule compression ratio reflects the impact of delays on manpower requirements. For example, if a transmission line project was originally scheduled to be completed in 30 days, but an upstream substation delay leaves only 20 days available, the schedule compression ratio is 20 divided by 30, or 0.67. This means that the same amount of work must be completed in a shorter time, resulting in a corresponding increase in manpower requirements. A task originally scheduled for 10 linemen would require 10 divided by 0.67, or approximately 15, in an emergency. This calculation assumes that the total workload remains unchanged and estimates the required additional manpower by compressing the schedule.

[0038] It should be noted that the process of generating the personnel supply-demand gap table reflects a refined management approach. By comparing the number of personnel required under emergency conditions with the actual number of personnel on staff, we can accurately identify the supply and demand situation for each job type and skill level. A positive value indicates a shortage of personnel in that position and the need for staffing; a negative value indicates a surplus of personnel who can be redeployed to other positions. This gap calculation provides a quantitative basis for subsequent staffing decisions. Urgency values ​​are assigned using a graded quantitative approach. Project delays of 1-3 days are assigned an urgency value of 1, delays of 4-7 days are assigned a value of 2, and delays of 8 days or more are assigned a value of 3. This grading reflects the varying impacts of different delays on the overall project schedule. A higher urgency level indicates a higher staffing priority for that position. Given limited resources, staffing needs for high-urgency positions should be prioritized. The construction of a two-dimensional supply-demand gap matrix integrates multidimensional information. Each element of the matrix consists of three components: the number of gaps reflects the extent of the staffing shortage, the number of surpluses indicates available resources, and the urgency level guides allocation priorities.

[0039] For example, the matrix elements for a senior electrician position might show a vacancy of 5, a surplus of 0, and an urgency of 3, indicating an urgent need for five senior electricians. Conversely, the elements for a junior civil engineer position might show a vacancy of 0, a surplus of 8, and an urgency of 1, suggesting that these eight junior civil engineers could support other in-demand positions through skills training or job adjustments. This matrix representation enables managers to comprehensively understand human resource allocation and optimize deployment across trades and skill levels.

[0040] S103. Analyze the supply and demand difference matrix and identify gaps. Compare the supply and demand values ​​of each job skill level to identify gaps. If the gap in high-skilled personnel exceeds the threshold, query the personnel skill database to determine the candidate set and the idle high-skilled personnel resource pool.

[0041] Each element in the supply-demand discrepancy matrix is ​​traversed to extract the number of vacancies for each type of work and skill level. Positions with skill levels greater than the preset advanced skill threshold are identified as high-skilled positions. The total number of vacancies for these high-skilled positions is calculated. If this total exceeds the preset personnel vacancies threshold, a cross-project deployment requirement identifier is generated, recording the type of vacancies and required skill level. Based on the vacancies and required skill level in the cross-project deployment requirement identifier, all matching personnel records are queried in the personnel skills database. The current task status field of each person is read, and personnel with a task status of completed or unassigned are identified as available personnel. The project affiliation information and estimated free time of these available personnel are obtained. Based on the project affiliation information of the available personnel, the project location coordinates are queried, and the deployment distance between them and the vacancies project is calculated. Personnel with a deployment distance less than the preset distance threshold are screened out, and the personnel are sorted from early to late according to their estimated free time to form a candidate set. From the candidate set, personnel whose task status is completed and whose estimated idle time is earlier than the start time of the gap project are extracted. The numbers, skill levels, job types and available time intervals of these personnel are summarized to build a resource pool of idle high-skilled personnel.

[0042] Specifically, traversal analysis of the supply-demand discrepancy matrix is ​​a key step in optimizing human resource allocation. This matrix is ​​constructed with job types as rows and skill levels as columns, with each element containing information on three dimensions: number of vacancies, number of redundancies, and degree of urgency.

[0043] Specifically, the preset high-skill threshold is typically set at skill level 7 or above. This standard is based on the power industry's skill certification system, where levels 1-3 are junior workers, levels 4-6 are intermediate workers, levels 7-9 are senior workers, and level 10 is a special technician. When the cumulative number of vacancies for a specific high-skill position exceeds the threshold of five, it means that relying solely on internal project deployment can no longer meet the demand, and a cross-project deployment mechanism must be activated.

[0044] In one possible implementation, the generation of the cross-project deployment requirement identifier includes detailed gap information.

[0045] For example, when a 220kV substation project has a shortage of eight senior electricians, a generated identifier records not only the number of vacancies but also includes the job code "DG," skill level requirement "L7," and project urgency "U3." This coded information provides precise search criteria for subsequent database queries, avoiding the efficiency loss caused by fuzzy matching. The personnel skills database is a core asset for enterprise human resources management, storing all employee skill certification information, project history, and current status. The task status field uses a standardized code: "01" indicates a task in progress, "02" indicates completed, and "03" indicates unassigned. When searching for senior electricians, the system retrieves all records of individuals holding a Level 7 electrician certificate or above and with a task status of "02" or "03." Estimated idle time is calculated based on the project schedule. For individuals with completed tasks, idle time starts at the current time. For individuals with ongoing tasks, completion time is estimated based on the remaining workload and standard working hours.

[0046] It's important to note that the calculation of deployment distances takes into account not only geographic location but also transportation accessibility. Project location coordinates are typically expressed in latitude and longitude, and the straight-line distance between two points is calculated using the spherical distance formula. The preset distance threshold is generally set at 200 kilometers, a value that takes into account the time cost and impact on the lives of personnel transfers. Deployments exceeding this distance often lead to increased employee resistance, impacting work efficiency. Therefore, when forming a candidate set, prioritizing personnel with closer proximity reduces deployment costs and increases employee acceptance. Sorting the candidate set based on expected idle time has important practical implications. Personnel who become idle earlier can be deployed to new projects more quickly, reducing waiting times.

[0047] For example, if Senior Electrician A will complete their current task in three days, while Electrician B will need ten days, Electrician A should clearly be prioritized for urgent project needs. This time-based consideration ensures efficient utilization of human resources. The establishment of a resource pool of available highly skilled personnel enables centralized resource management and rapid deployment. Each record in the resource pool contains complete deployment information: the personnel number for unique identification, the skill level to ensure capability matching, the job type to ensure professional matching, and the deployment window to clearly define the personnel's availability.

[0048] For example, senior electrician Zhang, with ID "E2024001" and skill level 8, is available for deployment from March 15, 2024, to April 30, 2024. This structured information organization enables managers to quickly match demand with resources, making deployment decisions within minutes and significantly improving response times to project delays.

[0049] S104: Recalculate the earliest start time and latest completion time of each operation node in the affected downstream projects according to the project dependency graph and resource constraints to obtain an optimized time window distribution plan.

[0050] The set of downstream project nodes affected by the delay is extracted from the project dependency graph. The originally planned start time, duration, and predecessor node number of each node are read. The earliest possible start time of the current node is updated based on the actual completion time of the predecessor node. The larger of the earliest possible start time and the originally planned start time is used as the adjusted start time. Based on the adjusted start time and duration of each node, the estimated operating period for each node is calculated. The number of personnel and equipment required for each node during the operating period is queried from the resource requirement table. The resource requirements of all nodes within the same period are accumulated and compared with the total available resources for that period in the resource supply database to identify periods where resource demand exceeds supply and the amount of excess resources. For periods with excess resources, all operating nodes within that period are extracted, and the longest path time from each node to the project endpoint is calculated as the time margin. The nodes are sorted from smallest to largest by time margin. For nodes with larger time margins, the operating period is reallocated between the completion time of their predecessor task and the start time of the subsequent task to avoid periods of excess resources. According to the reallocated job period, the start time and completion time boundaries of each node are updated. The interval between the earliest value of the start time and the latest value of the completion time is defined as the time window of the node. The time window information of all nodes is summarized to obtain the optimized time window distribution plan.

[0051] Specifically, the project dependency graph is a core data structure in power grid project management. It represents the sequential relationship between project nodes in the form of a directed graph. Each node represents a specific project task, such as substation construction, equipment installation, or line installation. Directed edges between nodes represent pre-depen- dencies. When an upstream project is delayed, all downstream nodes that depend on it are affected.

[0052] For example, the installation of the main transformer of a 220kV substation can only begin after the civil construction of the substation is completed. If the civil construction is delayed by 10 days, the earliest possible start time of the main transformer installation will also be postponed by 10 days.

[0053] Specifically, the adjusted start time is determined using a comparative selection mechanism. Two time points are considered simultaneously: the earliest possible start time calculated based on the predecessor's actual completion time, and the originally scheduled start time of the node. The larger of the two is used as the adjusted start time. This approach considers the hard constraints of dependencies while retaining the buffer time in the original plan.

[0054] For example, if a line construction task was originally scheduled to begin on March 15th, but the prerequisite foundation construction task was not completed until March 20th, the line construction would be adjusted to start on March 20th. Accumulating resource requirements is a key step in identifying resource conflicts. The resource requirement table details the various resources required for each project node, including the number of workers for different types of work and the number of equipment such as cranes and excavators. When the work periods of multiple project nodes overlap, the resource requirements for the same period need to be accumulated.

[0055] For example, from April 1 to April 5, the installation of substation equipment required 8 senior electricians and 2 cranes, and the simultaneous line installation required 5 senior electricians and 1 crane. The cumulative demand for this period was 13 senior electricians and 3 cranes.

[0056] In one possible implementation, the time slack calculation is based on path analysis of the network graph. Starting from the current node, all possible paths are followed to the project endpoint. The duration of all subsequent tasks along each path is calculated. The longest path duration is the time slack for that node.

[0057]

[0058] , T i represents the time margin of node i, P i represents the set of all possible paths from node i to the end point of the project, R ik represents the set of all subsequent tasks starting from node i on path k, d j represents the duration of task j. The max function is used to find the path with the longest total duration among all paths. The smaller the time slack, the greater the impact of the node on the overall duration and the more limited the room for adjustment. A node with a time slack of zero is on the critical path, and any delay will directly affect the total project duration.

[0059] It's important to note that the identification and handling of resource overage periods embodies the principle of optimization under resource constraints. When resource demand exceeds available supply during a period, the system prioritizes ensuring the on-time execution of critical tasks with smaller time margins, while adjusting the timeframe for non-critical tasks with larger time margins. This adjustment occurs within the feasible interval between the completion time of the predecessor task and the start time of the successor task.

[0060] For example, if the predecessor task of a certain equipment debugging task is completed on March 10, and the subsequent task is required to complete the debugging before March 25, then the task can be scheduled at any time between March 10 and March 25 to avoid the resource shortage period. The final determination of the time window combines multiple constraints. The time window of each node is defined by the earliest value of the start time and the latest value of the completion time. This interval not only satisfies the project dependency but also takes into account resource availability. The optimized time window distribution scheme not only provides a feasible operation range for each node, but also alleviates resource conflicts through reasonable time mismatching, realizes the optimization of the construction period under the conditions of limited resources, and effectively reduces the risk of secondary delays caused by resource shortages.

[0061] S105. Calculate the matching score between the personnel skill level and the engineering requirement characteristics by combining the personnel candidate set and the time window distribution scheme. If the matching score is higher than the set standard value, generate a personnel deployment recommendation list.

[0062] Extract the skill level value, professional certification category, and available time interval of each candidate from the candidate set. Read the operation period and task technical requirements of each project node from the time window distribution plan. Calculate the number of days of overlap between the personnel's available time interval and the project operation period. Divide the number of overlapping days by the total number of operation period days to obtain the time matching ratio. Filter out feasible personnel based on the time matching ratio being greater than the preset time threshold. For each feasible personnel, subtract their skill level value from the minimum skill level required for the project task. If the difference is greater than zero, it is used as the skill advantage value. If it is less than or equal to zero, it is set to zero. Compare the professional certification category with the required profession for the task. If there is a complete consistency, the professional compliance value is 1, and if there is a partial correlation, the value is 0.8. Use the preset weight parameters to calculate the comprehensive matching score. Multiply the skill advantage value by 0.3, the professional compliance by 0.4, and the time matching ratio by 0.3. Add the three values ​​to obtain a matching score between 0 and 1, S = 0.3 × A. skill +0.4×M prof +0.3×R time , S represents the comprehensive matching score, A skill Indicates the skill advantage value, M prof Indicates professional compliance, R time Represents the time matching ratio. This formula calculates the weighted sum of the three dimensions using preset weights, with professional conformity having the highest weight of 0.4, and skill advantage value and time matching ratio both having a weight of 0.3, resulting in a matching score between 0 and 1. If the score is higher than the set standard value of 0.7, the person and their score are recorded in the list of compatible personnel. From the list of compatible personnel, the matching scores are sorted from high to low, and each person's number, current project, recommended deployment start and end time, and target project node number are extracted to form a list of recommended personnel deployments.

[0063] Specifically, calculating the time matching ratio is the first step in assessing the feasibility of staffing deployment. The staffing candidate set records each candidate's available time interval, which represents the time between the completion of the current task and the start of the next scheduled task.

[0064] For example, a senior electrician's available time is from April 10th to April 25th, a total of 16 days. The target project node's work period is from April 15th to April 30th, a total of 16 days. The two time periods overlap from April 15th to April 25th, for a total of 11 days. The time matching ratio is 11 divided by 16, which is approximately 0.69. This ratio reflects the degree of time coverage available for the person to participate in the target project.

[0065] Specifically, the preset time threshold is typically set at 0.6, meaning that personnel must cover at least 60% of the project's operating time to be considered time-feasible. This threshold is based on project continuity requirements. Too low a time coverage ratio can lead to frequent personnel turnover, impacting work efficiency and quality. Calculating the time matching ratio lays the foundation for subsequent comprehensive evaluations, ensuring the time-based feasibility of the deployment plan. The skill advantage value reflects the degree of match between personnel capabilities and task requirements. Power engineering employs a ten-level skill level system, with Level 7 and above considered advanced. If a substation equipment commissioning task requires a minimum Level 7 skill level, and the candidate holds a Level 9 certificate, the skill advantage value is 2. This value not only indicates that the individual meets the basic requirements but also reflects the depth of their skill reserves. A higher skill advantage value indicates that the individual can handle complex situations that may arise during the task, improving the reliability of project implementation.

[0066] In one possible implementation, professional compatibility is determined using a tiered evaluation method. Complete alignment means the candidate's professional certification category perfectly matches the task requirements. For example, if the task requires "High Voltage Electrical Testing," the candidate happens to hold a certificate in that field. Partial correlation includes cases with similar expertise. For example, if the task requires "Relay Protection Commissioning," the candidate holds a certificate in "Electrical Secondary Circuit Commissioning." Since there's significant overlap in skill requirements, a score of 0.8 is assigned. This tiered approach prioritizes matching professionals while also allowing for the transfer of similar expertise.

[0067] It's important to note that the weighting parameters reflect the relative importance of different dimensions in the match assessment. Professional fit is weighted the highest at 0.4, as mismatching professional skills directly impacts task quality. Skill advantage and time fit each account for 0.3, reflecting a balanced consideration of ability and time availability. The match score calculation process ensures the results fall within a standardized range of 0 to 1, facilitating consistent comparison. The standard value of 0.7 means that candidates must perform well in all three dimensions to be selected. The generation and sorting of the list of suitable candidates provides a clear basis for final decision-making. Sorting by match score, from highest to lowest, allows managers to prioritize the most suitable candidates. The key information included in the recommended personnel transfer list ensures the feasibility of transfer instructions: the personnel number provides precise location information, current project information facilitates coordination of transfer procedures, the recommended start and end times clarify the timeframe for the personnel's involvement, and the target project node number indicates the specific destination. This structured information organization significantly improves the efficiency and accuracy of cross-project personnel transfers.

[0068] S106. Calculate the shortest deployment path and time cost for cross-project personnel flow based on the recommended list of personnel deployment, and determine the target personnel flow plan based on the transportation distance between project sites, deployment cost, skill matching degree, and idle high-skilled personnel resource pool.

[0069] Based on the current project location and target project node location of each candidate in the proposed personnel deployment list, the project site location database is queried to obtain the latitude and longitude coordinates of each site. The Dijkstra algorithm is used to calculate the shortest travel distance from each candidate's current location to the target location via the road network. This distance is divided by the preset average driving speed to obtain the travel time. Based on the travel time and the shortest travel distance, the per-kilometer transportation cost and the daily personnel deployment subsidy amount are retrieved from the preset cost parameter table. The transportation cost is calculated by multiplying the shortest travel distance by the per-kilometer transportation cost. The travel time, converted into days, is multiplied by the daily personnel deployment subsidy amount to obtain the time subsidy. The two are added together to obtain the individual personnel deployment cost. For each candidate's deployment cost, the maximum deployment cost of all candidates is calculated. The current deployment cost of the candidate is subtracted from the maximum value and then divided by the maximum value to obtain the cost advantage score. The skill match score of the candidate is retrieved from the proposed list. The cost advantage score is multiplied by 0.4 and the skill match score is multiplied by 0.6 to obtain the overall evaluation value. All candidate personnel are sorted from high to low according to the comprehensive evaluation value, and the latest status of the top-ranked personnel in the idle high-skilled personnel resource pool is checked in turn. Personnel with available status are selected until the number of personnel required for the project is met. The number, deployment route, expected arrival time and respective deployment cost of the selected personnel are summarized to form a target personnel flow plan.

[0070] Specifically, the application of the Dijkstra algorithm in power grid project personnel deployment demonstrates the importance of path optimization. The algorithm gradually expands the shortest path tree to find the shortest path from a starting point to each node. In practice, a project site location database stores the precise geographic coordinates of all ongoing and pending projects. These coordinates form the nodes of a road network graph. Road connectivity and mileage information within the highway network form the edges between nodes.

[0071] For example, there may be multiple routes to choose from substation A to transmission line construction site B. The algorithm will comprehensively consider the actual mileage of each section of road and calculate the route with the shortest total distance, such as a route of 156 kilometers via a national highway and then a provincial highway, rather than a 180-kilometer route via a highway.

[0072] Specifically, the calculation of travel time needs to take into account the differences in driving speeds on different types of roads. The preset average driving speed is usually set according to the road grade classification, 80 km / h for expressways, 60 km / h for national highways, and 40 km / h for provincial highways. This classification setting is closer to the actual situation and makes the time estimate more accurate. A 156-kilometer mixed road section may include 50 kilometers of expressways, 80 kilometers of national highways, and 26 kilometers of provincial highways, with corresponding driving times of 0.625 hours, 1.33 hours, and 0.65 hours, respectively, for a total of approximately 2.6 hours. The setting of the cost parameter table reflects the economic considerations of personnel deployment. The per-kilometer transportation cost includes comprehensive costs such as vehicle fuel, tolls, and vehicle depreciation, and is usually set at 2-3 yuan / kilometer. The daily personnel deployment allowance takes into account the additional living costs incurred by personnel leaving their original base, and is generally 200-300 yuan / day.

[0073] For example, a senior electrician is transferred from place A to place B, which is 200 kilometers away. The journey is expected to take 4 hours. The transportation cost is 200×2.5=500 yuan, and the time allowance is 0.5 day×250=125 yuan. The total transfer cost is 625 yuan.

[0074] In one possible implementation, the cost advantage score is calculated using a reverse normalization method. The core idea of ​​this method is to convert the negative cost metric into a positive score. Assuming the highest deployment cost among all candidates is 3,000 yuan, and the deployment cost of a particular candidate is 625 yuan, then their cost advantage score is (3,000 - 625) / 3,000 ≈ 0.79. A higher score indicates a more economical deployment. Compared to directly using the inverse of cost, this method avoids numerical anomalies, and the result is always between 0 and 1.

[0075] It's important to note that the weighting of the comprehensive evaluation scores reflects a balanced approach to decision-making. The cost advantage score is weighted 0.4, and the skill match score is weighted 0.6. This configuration slightly favors skill match, as power grid projects have high technical requirements and personnel competence directly impacts project quality and safety. However, cost considerations cannot be ignored, especially when deploying large-scale personnel, where cost control has a significant impact on the overall project effectiveness. The process of developing the target personnel flow plan embodies multi-objective optimization. By ranking by comprehensive evaluation scores, the system is able to find the optimal balance between skill match and cost control. The plan not only incorporates information on individual personnel deployments but also considers the synergistic effects of multiple personnel combinations.

[0076] For example, if a project requires three senior electricians, the system might select the first, third, and fifth ranked personnel in comprehensive evaluation, because the second and fourth personnel are from the same project and their simultaneous transfer would affect the original project's operations. The plan details each personnel's deployment route, estimated arrival time, and cost, providing project managers with a comprehensive basis for implementation.

[0077] Extract the coordinates of the candidate personnel's work locations from the personnel deployment recommendation list, calculate the straight-line distance and actual traffic route between the project sites, combine the matching scores of personnel's professional skill certification levels with the target engineering and technical requirements, query the number of available personnel and skill distribution in the idle high-skilled personnel resource pool, combine the deployment distance cost, skill adaptability and personnel availability, screen personnel combinations based on examples and skill matching, and form a cross-project personnel flow deployment plan.

[0078] Extract the latitude and longitude coordinates of each candidate's current work location from the proposed personnel deployment list. Obtain the geographic coordinates of the target project from the project location database. Calculate the straight-line distance between two points using the spherical distance formula: d = R·arccos(sinφ1sinφ2+cosφ1cosφ2cos(λ2-λ1)), where d represents the spherical distance between the two points, R represents the radius of the sphere, φ1 represents the latitude of the first point, φ2 represents the latitude of the second point, λ1 represents the longitude of the first point, λ2 represents the longitude of the second point, and arccos represents the inverse cosine function. Query the traffic network database to obtain the mileage of the actual drivable route. Based on the mileage of the actual drivable route, query the preset transportation cost standard table to obtain the transportation fee per kilometer. Multiply the mileage by the transportation fee per kilometer to obtain the deployment distance cost for each individual. Simultaneously, read the candidate's professional skill certification level and the minimum level required for the target project. Subtract the minimum level from the minimum level to obtain the skill excess value. Based on skill excess values, the pool of available highly skilled personnel is searched. The number of available personnel with the same or higher skill excess values ​​is counted and sorted by skill level from high to low to form a sequence of available personnel. The distance cost ratio is calculated by dividing each candidate's deployment distance cost by the maximum distance cost of all candidates. The reciprocal of the distance cost ratio, multiplied by the preset distance weight, plus the skill excess value multiplied by the preset skill weight, is used to determine the personnel deployment priority score. Personnel are selected from the highest to the lowest priority score until the target project's required number of positions at each skill level is met. The selected personnel's number, origin, skill level, and deployment cost are summarized to form a cross-project personnel mobility deployment plan.

[0079] Specifically, the spherical distance formula, used in power grid project personnel deployment, takes into account the impact of the Earth's curvature on long-distance measurements. When two projects are far apart, simple planar distance calculations can produce significant errors. Spherical distance is based on the longitude and latitude coordinates of two points and uses the principle of great-circle navigation to calculate the shortest distance.

[0080] For example, the spherical distance from the substation in City A at 30.5 degrees north latitude and 114.3 degrees east longitude to the power transmission project in City B at 31.2 degrees north latitude and 121.5 degrees east longitude is about 830 kilometers, while the plane projection distance may reach more than 850 kilometers. This 20-kilometer difference will directly affect the accuracy of the cost calculation.

[0081] Specifically, actual drivable routes are often longer than straight-line distances because they must navigate mountains, rivers, and follow existing transportation networks. The transportation network database stores detailed information on major highways nationwide, including road grades, mileage markings, and connectivity. Path planning algorithms search for feasible routes within the road network and calculate actual mileage. D actualrepresents the actual mileage calculated by the path planning algorithm, n represents the total number of nodes in the path, x i 、y i 、z i They represent the three-dimensional coordinates of the i-th path node respectively. The formula calculates the total mileage of the actual highway by accumulating the Euclidean distances between adjacent path nodes.

[0082] The actual road distance between City A and City B is approximately 950 kilometers, 120 kilometers longer than the spherical distance. This discrepancy is more pronounced in mountainous areas or areas with dense waterways. The standard transportation cost table reflects the cost differences between different modes of transportation and distance ranges. Short-distance transportation costs are higher per kilometer because fixed costs are less widely distributed; long-distance transportation, however, tends to have lower unit costs due to economies of scale.

[0083] For example, for a transfer within 100 kilometers, the transportation fee might be 3.5 yuan per kilometer; for a transfer between 100 and 500 kilometers, it drops to 2.8 yuan per kilometer; and for transfers over 500 kilometers, it's 2.2 yuan per kilometer. For a transfer between City A and City B, a distance of 950 kilometers, the transportation cost is 950 x 2.2 = 2,090 yuan. This tiered pricing model better aligns with the actual economics of transportation.

[0084] In one possible implementation, the calculation of skill excess reflects the importance of personnel capacity reserves. The power industry's skill levels range from 1 to 10, with each level representing deeper expertise and richer practical experience. For example, if a 500kV substation maintenance task requires a skill level of 7, a Level 9 senior technician would have a skill excess of 2, meaning they are not only capable of performing basic tasks but also handling complex and unexpected faults. In contrast, a person who only meets the Level 7 requirement would have a skill excess of 0. While meeting the minimum requirement, their ability to handle special situations is limited.

[0085] It's important to note that querying and statistics on the pool of available highly skilled personnel provides a holistic perspective for decision-making. The resource pool is managed hierarchically by skill level, with personnel status updated in real time. When a Level 9 technician is needed, the system not only displays the number of Level 9 personnel but also lists the availability of Level 10 Special Technicians, as higher-level personnel are backward compatible. This structured sequence of available personnel ensures the feasibility and flexibility of deployment plans. The introduction of a distance cost ratio enables fair comparison across different distance scales. Assuming the farthest deployment distance among candidate candidates is 1,200 kilometers, and a candidate's deployment distance is 400 kilometers, their distance cost ratio is 400 / 1,200, which is ≈ 0.33. Taking the inverse of this is 3, indicating that this candidate's distance advantage is three times that of the farthest candidate. Combining the preset distance weight of 0.4 and skill weight of 0.6, a candidate with a skill excess of 2 and a distance advantage of 3 has a deployment priority score of 3 × 0.4 + 2 × 0.6 = 2.4. This comprehensive evaluation method balances economic and technical considerations, and the resulting allocation plan not only ensures project quality but also controls allocation costs.

[0086] S107. Update the personnel allocation table and skill structure distribution of each project according to the target personnel flow plan, recalculate the project progress forecast value and resource utilization index, generate a dynamic personnel reallocation execution plan and update the system configuration parameters.

[0087] Based on the personnel transfer-in and transfer-out information in the target personnel flow plan, records of transferred-out personnel are deleted from each project's staffing table and records of transferred-in personnel are added to the target project's staffing table. The distribution of on-the-job headcount for each project is recalculated by job type and skill level to generate updated project staffing data. Based on this updated project staffing data, the workload of each project's unfinished tasks is extracted from the engineering task list. The average daily work efficiency values ​​corresponding to different skill levels are queried. The number of on-the-job personnel for each skill level is multiplied by the corresponding average daily work efficiency values ​​and added together to obtain the project's overall daily completion capacity. The unfinished workload is then divided by the overall daily completion capacity to obtain the estimated number of days required. The estimated number of days required is added to the current date to obtain the new estimated completion date. The staffing adequacy ratio is calculated as the ratio of the planned staffing headcount to the actual staffing headcount for each time period. The staffing execution records, the new estimated completion date, and the staffing adequacy ratio data are integrated to form a project status update dataset. The project status update data set is used to generate a dynamic personnel reallocation execution plan that includes personnel deployment details, project progress adjustment values, and resource allocation optimization parameters. The personnel deployment time, skill level distribution ratio, and project priority weight parameters in the plan are written into the system configuration file to complete the configuration parameter update.

[0088] Specifically, updating the project staffing table is fundamental to dynamic human resource management. This table stores personnel information for each project in a two-dimensional structure, with rows representing different job categories and columns representing skill level distribution. Once a staff turnover plan is finalized, the system must simultaneously update the configuration data for the relevant projects.

[0089] For example, if two Level 9 senior electricians from a 220kV substation project were deployed to support a delayed 500kV transmission line project, the system would decrement the value in the Level 9 column in the Electrician row for the substation project by 2, while simultaneously incrementing the corresponding value in the transmission line project by 2. This real-time update mechanism ensures accurate and consistent personnel statistics.

[0090] Specifically, the engineering task list records all work package information after the project is broken down, including task name, workload estimate, and completion status. The workload of unfinished tasks is usually measured in standard working hours. For example, the remaining construction workload of a 10-kilometer section of transmission line is 800 standard working hours. The average daily work efficiency values ​​of personnel with different skill levels reflect their ability differences. A level 7 technician can complete 8 standard working hours per day, a level 8 technician can complete 10 working hours, and a level 9 technician can complete 12 working hours. This differentiated setting is based on a large amount of historical data statistics and reflects the actual impact of skill level on work efficiency.

[0091] In one possible implementation, the calculation of overall daily completion capacity requires comprehensive consideration of the team's skill mix. Assuming the adjusted project team consists of five Level 7 technicians, three Level 8 technicians, and two Level 9 technicians, the daily completion capacity is 5 × 8 + 3 × 10 + 2 × 12 = 94 standard working hours. For the remaining workload of 800 working hours, the estimated completion time is 800 ÷ 94, which is approximately 8.5 days, rounded up to 9 days. This calculation method is more realistic than a simple average because it fully accounts for the contributions of different skill levels.

[0092] It's important to note that the introduction of the staffing adequacy ratio provides a quantitative basis for project risk assessment. This metric is derived by comparing planned staffing with actual staffing numbers. A staffing adequacy ratio below 0.8 indicates a project faces a risk of human resource shortages.

[0093] For example, a project planned to require 20 people in the third week, but only 15 were actually on staff, with a staffing ratio of 0.75, indicating a potential schedule delay. This early warning mechanism enables managers to take proactive supplementary measures. The integration of project status update datasets enables the integrated management of multi-dimensional information. The dataset not only contains static staffing information but also dynamic schedule forecasts and resource utilization. Timestamps enable the system to track the history of each adjustment, creating a complete change trajectory. This data organization provides a solid foundation for subsequent trend analysis and decision optimization. Configuration file parameter updates reflect standardized plan execution requirements. Personnel deployment time parameters ensure orderly deployment, avoiding disruptions caused by large simultaneous staff turnover. Skill level distribution parameters are used to maintain a reasonable project team structure, such as maintaining a 2:3:5 ratio of high, medium, and low-skilled personnel. Project priority weight parameters guide resource allocation, with key projects receiving higher weights. Systematically managing these parameters not only improves execution efficiency but also ensures scientific and traceable resource allocation, enabling dynamic optimization of human resources for power grid projects.

[0094] It will be apparent to those skilled in the art that the present application is not limited to the details of the exemplary embodiments described above and that the present application can be implemented in other specific forms without departing from the spirit or essential characteristics of the present application. Therefore, the embodiments should be considered in all respects as illustrative and non-restrictive, and the scope of the present application is defined by the appended claims, not the foregoing description, and all variations within the meaning and range of equivalents of the claims are intended to be included therein. Any reference sign in a claim should not be construed as limiting the claim to which it relates.

Claims

1. An enterprise resource scheduling optimization method based on artificial intelligence and large model algorithm, characterized in that: The method comprises: The system obtains the current progress and delay information of each power grid project, analyzes the completion status and remaining workload of the project nodes, and calculates the number of days of delay for each project and the number of affected downstream projects, power supply areas, and users based on standard construction period parameters to obtain a project dependency graph. Based on the number of days of delay and the affected scope, the quantity and skill level requirements of different types of work at each project node are calculated to generate a supply and demand difference matrix. The supply and demand difference matrix is ​​analyzed to identify the gap in high-skilled personnel and determine a candidate set and a resource pool of idle high-skilled personnel. Based on the project dependency graph and resource constraints, the earliest start time and latest completion time of each work node are recalculated to obtain a time window distribution plan. Based on the candidate set and the time window distribution plan, the matching score between the personnel skill level and the project demand characteristics is calculated to generate a personnel deployment recommendation list. Based on the personnel deployment recommendation list, the cross-project personnel flow path and time cost are calculated, and the target personnel flow plan is determined based on the transportation distance between project sites and the degree of skill matching. The project progress forecast value is recalculated based on the target personnel flow plan, a dynamic personnel reallocation execution plan is generated, and the system configuration parameters are updated.

2. The enterprise resource scheduling optimization method based on artificial intelligence and large model algorithm according to claim 1 is characterized in that: The current progress and delay information of each power grid project is obtained, the completion status of the project nodes and the remaining workload are analyzed, and the number of days of delay for each project and the number of affected downstream projects, power supply areas, and users are calculated in combination with standard construction period parameters to obtain a project dependency diagram, including: Obtain the planned start time, planned end time and actual completion time of each power grid engineering project, read the completion status identifier of each node from the engineering node list, and calculate the ratio of the number of completed nodes to the total number of nodes to obtain the progress rate; query the standard working hours corresponding to each unfinished node from the engineering quota database, and accumulate the standard working hours to obtain the remaining work; calculate the difference between the actual completion time and the planned end time to determine the number of extension days; query the predecessor and successor relationship data between projects according to the project number, extract the downstream project list, and count the number of projects in the downstream project list; query the power supply facility code according to the project number in the downstream project list, determine the affected power supply area code set through the mapping relationship between the power supply facility and the power supply area, query the user archive database, and accumulate the total number of users in each area to obtain the number of affected users; construct a directed graph based on the predecessor and successor relationship, mark the number of extension days, the number of affected downstream projects, the power supply area code set and the number of users on the nodes, and the edges represent the dependency relationship, and generate the engineering dependency graph through the depth-first traversal algorithm.

3. The enterprise resource scheduling optimization method based on artificial intelligence and large model algorithm according to claim 1 is characterized in that: According to the number of days of delay and the scope of impact, the quantity demand and skill level requirements of different types of work at each project node are calculated to generate a supply and demand difference matrix, including: According to the number of extended days, the construction period adjustment parameter table is queried to obtain the construction period compression ratio and calculate the compressed construction period; the task list and work type identification of each engineering node are extracted from the affected range, the work type personnel configuration standard table is queried to obtain the standard number of personnel for each work type required for unit task volume, and the number of required personnel under the expedited state is calculated in combination with the task intensity coefficient; the technical difficulty level of the task is read from the task list, the work type skill requirement comparison table is queried to determine the minimum skill level and construct a personnel demand matrix; the current list of available personnel is obtained, the work type and skill level certificate number are read and a personnel supply matrix is ​​constructed; the corresponding elements of the personnel demand matrix and the personnel supply matrix are subtracted to obtain a supply and demand difference matrix, the number of high-skilled personnel gaps and the number of low-skilled personnel redundancies are identified and the supply and demand difference matrix is ​​constructed.

4. The enterprise resource scheduling optimization method based on artificial intelligence and large model algorithm according to claim 1 is characterized in that: Generating the supply-demand difference matrix includes: According to the type of work identification of each task in the task list, the correspondence table between engineering task type and skill level is queried to determine the minimum skill level required for each task; the list of on-the-job personnel is extracted from the personnel attendance data, the skill certification database is queried, and the number of on-the-job personnel of each type of work and each skill level is counted; according to the number of delay days and the construction period compression ratio, the number of required personnel of each type of work and each skill level under emergency conditions is calculated, and subtracted from the number of on-the-job personnel to obtain a personnel supply and demand difference table; according to the personnel supply and demand difference table, the number of vacancies in high-skilled positions and the number of redundancies in low-skilled positions are identified, and a supply and demand difference matrix containing the number of vacancies and the number of redundancies is constructed.

5. The enterprise resource scheduling optimization method based on artificial intelligence and large model algorithm according to claim 1 is characterized in that: The analyzing the supply-demand difference matrix, identifying the gap in high-skilled personnel, and determining a candidate set and an idle high-skilled personnel resource pool includes: Traverse the supply-demand difference matrix to extract the total number of vacancies in high-skilled positions; query the personnel skills database based on the vacancies’ job types and required skill levels to obtain the task status and project affiliation information of the deployable personnel; calculate the deployment distance between the deployable personnel and the vacancies’ projects, filter out personnel with a distance less than a threshold, and sort them to form the candidate set; extract personnel whose task status is completed and whose idle time is earlier than the start time of the vacancies’ projects from the candidate set, summarize the number, skill level, job type, and available deployment time, and construct the idle high-skilled personnel resource pool.

6. The enterprise resource scheduling optimization method based on artificial intelligence and large model algorithm according to claim 1 is characterized in that: The earliest start time and latest completion time of each job node are recalculated according to the project dependency graph and resource constraints to obtain a time window distribution plan, including: Extract the set of downstream engineering nodes from the engineering dependency graph, read the original planned start time and the actual completion time of the predecessor node, and determine the adjusted start time; calculate the expected operation period of each node, summarize the resource requirements, compare with the resource supply database, and identify the resource excess period; sort the nodes in the resource excess period according to the time margin, and reallocate the operation period; update the start time and completion time boundary of each node, and summarize to form the time window distribution plan.

7. The enterprise resource scheduling optimization method based on artificial intelligence and large model algorithm according to claim 1 is characterized in that: The step of combining the candidate set and the time window distribution scheme to calculate a matching score between personnel skill levels and engineering requirement characteristics and generating a personnel deployment recommendation list includes: The skill level values ​​and available deployment time of the candidate personnel are extracted from the candidate set, the engineering node operation period and task technical requirements are read from the time window distribution plan, and the time matching ratio is calculated; the time feasible personnel are screened, and the skill advantage value and professional conformity are calculated; the matching score is calculated, and the personnel with scores higher than the standard value are screened, and the personnel number, current project, deployment time and target engineering node are summarized after sorting to form the personnel deployment recommendation list.

8. The enterprise resource scheduling optimization method based on artificial intelligence and large model algorithm according to claim 1 is characterized in that: Based on the personnel deployment suggestion list, calculate the cross-project personnel flow path and time cost, and determine the target personnel flow plan based on the transportation distance and skill matching between project sites, including: The current location of the personnel and the location of the target project node are obtained from the personnel deployment suggestion list, the longitude and latitude of the construction site are queried, and the shortest transportation distance and time are calculated; the deployment cost is calculated based on the transportation distance and time; the comprehensive evaluation value is calculated by combining the deployment cost and the skill matching degree score; the personnel are sorted by the comprehensive evaluation value, the status of the personnel in the idle high-skilled personnel resource pool is checked, the deployable personnel are selected, the number, path, arrival time and cost are summarized, and the target personnel flow plan is formed.

9. The enterprise resource scheduling optimization method based on artificial intelligence and large model algorithm according to claim 1 is characterized in that: The above calculation of cross-project personnel flow paths and time costs, combined with the transportation distance between project sites and the degree of skill matching, determines the target personnel flow plan, including: Extract the candidate personnel's work location coordinates and target project coordinates from the personnel deployment suggestion list, calculate the straight-line distance and actual path mileage; calculate the deployment distance cost based on the actual path mileage; read the candidate personnel's skill level value and calculate the skill excess value; query the idle high-skilled personnel resource pool based on the skill excess value, and sort to form an available personnel sequence; combine the distance cost ratio and skill excess value to calculate the priority score, select personnel, summarize the number, origin, skill level and deployment cost, and form a cross-project personnel flow deployment plan.

10. The enterprise resource scheduling optimization method based on artificial intelligence and large model algorithm according to claim 1 is characterized in that: The recalculating the project progress forecast value according to the target personnel flow plan, generating a dynamic personnel reallocation execution plan and updating system configuration parameters includes: According to the target personnel flow plan, the project personnel allocation table is updated and the distribution of the number of people on duty in each project is counted; the workload of unfinished tasks is extracted from the engineering task list, and the estimated number of days and completion date are calculated based on the average daily work efficiency of the skill level; the personnel adequacy rate is calculated, and the deployment records, completion date and adequacy rate are integrated to form a project status update data set; based on the project status update data set, a dynamic personnel reallocation execution plan containing deployment details and progress adjustment values ​​is generated, and the deployment time and skill level distribution ratio in the system configuration file are updated.

Images