Super-large scale network operation and maintenance method and system based on nested block inheritance

Abstract

The invention discloses a super-large scale network operation and maintenance method and system based on nested block heredity, and the method comprises the steps: obtaining asset network data, and constructing an infrastructure operation and maintenance planning physical mathematical model; performing annual plan exchange crossover of constraint perception on parent individuals representing different complete cycle planning schemes; blocking and nesting assets; maintaining strategy exchange and hierarchy variation among the blocks; and adaptive optimization guided by block-level linear programming. The system comprises a physical modeling module, an annual plan exchange cross module, a block nesting module, a sub-block processing module and a self-adaptive optimization module. By using the method, the calculation bottleneck under a huge asset scale can be solved, and an efficient and extensible decision support scheme is provided for fine management of complex infrastructure systems such as roads, bridges and the like. The method can be widely applied to the technical field of infrastructure management.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of infrastructure management, and particularly relates to a super-large-scale network operation and maintenance method and system based on nested block genetic algorithm. BACKGROUND

[0002] Infrastructure maintenance and repair (M&R) planning is the core of asset management system, whose core goal is to maintain the service level of infrastructure such as roads, bridges and water systems by optimizing intervention strategies. The existing technology usually models this problem as a stochastic integer programming or mixed integer nonlinear programming problem. However, when facing super-large-scale networks containing tens of thousands or even millions of assets in the real world, due to the large number of assets, various processing schemes and long planning period, the number of decision variables increases exponentially, causing the "combinatorial explosion" problem. This further leads to the extremely low computational efficiency of existing methods, which is difficult to meet the actual planning needs. SUMMARY

[0003] Therefore, in order to solve the technical problem of low computational efficiency caused by "combinatorial explosion" when the existing technology deals with large-scale infrastructure operation and maintenance planning, in the first aspect, the present application provides a super-large-scale network operation and maintenance method based on nested block genetic algorithm, which comprises the following steps: Physical network data modeling and quantification: obtaining asset network data and constructing a physical mathematical model of infrastructure operation and maintenance planning.

[0004] Generating and exchanging annual engineering plans: under the genetic algorithm framework, the parent individuals representing different complete cycle planning schemes are subjected to constraint-aware annual plan exchange and crossover.

[0005] Asset block and nesting: for annual plans that do not meet the constraints, the large-scale physical asset network is divided into blocks according to administrative divisions, road grades or natural geographical boundaries; if the number of assets in the block is still large, it is further recursively divided into smaller "sub-blocks", forming a multi-level nested structure, so as to decompose the original super-large-scale problem into multiple sub-problems with lower dimensions and easy management in the physical space.

[0006] Block maintenance strategy exchange and hierarchical variation: cross-operation is performed between sub-blocks to exchange asset subsets and promote global propagation of high-quality solution characteristics; at the same time, controlled maintenance, repair and reconstruction variation is introduced at each level of block.

[0007] Block-level linear programming guided adaptive optimization: for the identified infeasible sub-blocks, a part of them is selected for optimization through an adaptive capture mechanism. The linear programming (LP) is used to solve the conditionally optimal solution under the adjusted local constraints, efficiently repairing the infeasible solution and improving the quality of the local solution.

[0008] In addition to the above-mentioned method and process, the present invention also proposes a large-scale network operation and maintenance system based on nested block genetics. The system includes a physical modeling module, an annual plan exchange and cross-connection module, a block nesting module, a sub-block processing module, and an adaptive optimization module.

[0009] Based on the above scheme, this invention provides a method and system for ultra-large-scale network operation and maintenance based on nested block genetics. It successfully decomposes an ultra-large-scale, nonlinear, multi-constraint physical infrastructure asset network planning problem into a series of manageable and parallel-solvable sub-problems. Combined with an efficient local linear programming solver, it significantly improves computational efficiency and the quality of the final planning scheme while ensuring the feasibility of the engineering scheme. Attached Figure Description

[0010] Figure 1 This is a flowchart illustrating the ultra-large-scale network operation and maintenance method based on nested block genetics proposed in this invention. Figure 2 This is a schematic diagram of constraint-aware annual plan exchange crossover in a specific embodiment of the present invention.

[0011] Figure 3 This is a schematic diagram illustrating asset segmentation and nesting in a specific embodiment of the present invention.

[0012] Figure 4 This is a schematic diagram of block cross-operation in a specific embodiment of the present invention.

[0013] Figure 5 This is a schematic diagram of adaptive optimization guided by block-level linear programming in a specific embodiment of the present invention.

[0014] Figure 6 This is an example diagram illustrating the road network planning results of a large city according to a specific embodiment of the present invention. Detailed Implementation

[0015] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0016] It should be noted that, for ease of description, only the parts relevant to the invention are shown in the accompanying drawings. Unless otherwise specified, the embodiments and features described in this application can be combined with each other.

[0017] It should be understood that the terms "system," "apparatus," "unit," and / or "module" used in this application are a method of distinguishing different components, elements, parts, sections, or assemblies at different levels. However, if other terms can achieve the same purpose, they may be replaced by other expressions.

[0018] As indicated in this application and claims, unless the context clearly indicates otherwise, the words "a," "an," "a," and / or "the" are not specifically singular and may include the plural. Generally, the terms "comprising" and "including" only indicate the inclusion of expressly identified steps and elements, which do not constitute an exclusive list, and the method or apparatus may also include other steps or elements. An element defined by the phrase "comprising an..." does not exclude the presence of other identical elements in the process, method, product, or apparatus that includes the element.

[0019] In the description of the embodiments of this application, "a plurality of" refers to two or more. The terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature.

[0020] Furthermore, flowcharts are used in this application to illustrate the operations performed by the system according to embodiments of this application. It should be understood that the preceding or following operations are not necessarily performed precisely in sequence. Instead, the steps can be processed in reverse order or simultaneously. Additionally, other operations can be added to these processes, or one or more steps can be removed from them.

[0021] Reference Figure 1 This is a flowchart illustrating an optional example of the large-scale network operation and maintenance method based on nested block genetics proposed in this invention. This method can be applied to computer devices, and the method proposed in this embodiment may include, but is not limited to, the following steps: Step S1: Based on the asset network data, construct an operation and maintenance planning model; For infrastructure networks, the decision-making problem involves planning the maintenance operations of n assets over the next h years, requiring the determination of specific maintenance, repair, and reconstruction (MRR) actions to be taken for each asset in the infrastructure network within each fiscal year. Specifically, this involves defining decision variables, objective functions, and constraints. The basic content includes: Decision variables. Indicates the first [number] in the planning period h. Year( ) for the first Individual assets ( The first step taken Maintenance, Repair and Reconstruction (MRR) Operations ).in =0 means no action is taken.

[0022] Objective function: Maximize the time-averaged service level of the network during the planning period. ), which is the weighted average service level for each year ( The mean of ).

[0023] in, For assets Weights (such as asset length, grade, etc.); For assets In the Year-end performance status depends on initial conditions. Historical maintenance, repair and reconstruction decisions and model parameters It is determined by the state transition model.

[0024] Specifically, state transition models can be deterministic or stochastic, depending on the specific characteristics of the assets they apply to. Their general mathematical representation is as follows: Constraints. The network prioritization planning problem must meet several constraints to ensure that the solution is within the available budget, maintains a certain level of service each year, and meets other technical and regulatory requirements.

[0025] Annual budget constraints: Ensure that the budget allocated to maintenance operations each year does not exceed the prescribed lower and upper limits.

[0026] in, Indicates the first Annual assets Take maintenance, repair and reconstruction actions The cost; Indicates the first The designated budget for the year; (0≤) ≤1) represents the tolerance of the budget in year t.

[0027] Total budget constraint: Limits the total cost of the maintenance plan over the entire planning period.

[0028] Among them, B total This indicates the total available budget for the planning year.

[0029] Performance constraints: Ensure that the system's service level remains above the specified threshold each year.

[0030] Mutually exclusive maintenance, repair, and rebuilding actions: Ensure that only one maintenance, repair, or rebuilding option is applied to each asset each year.

[0031] Step S2: Based on the asset network data, select a reference year and exchange the corresponding annual maintenance plans between the parent generations.

[0032] When crossing between parents, randomly select a complete year and perform the full-year decision operation between the two parents (i.e., that year). All corresponding assets of (Decision) Exchange; distinct from traditional random segment crossing.

[0033] Step S3: For infeasible annual maintenance plans, divide all assets within that year into a multi-level nested asset block structure. Here, "all assets" refers to all assets in the entire network that are involved in the operation and maintenance plan within a certain year, such as roads, bridges, pipelines, etc. in the entire infrastructure network (including asset status, weight, and operation and maintenance costs).

[0034] Step S4: Within a multi-level nested asset block structure, perform MRR mutation from top to bottom, and perform block crossover operations between child blocks under the same parent block. Step S5: Perform local optimization operations on the last sub-block that does not meet the budget constraints.

[0035] After completing the local optimization, the results of all sub-blocks are combined to form a revised and complete annual operation and maintenance plan.

[0036] In some feasible embodiments, step S2 specifically includes: The constraint-aware annual plan exchange method is specifically designed to maintain the financial feasibility of each annual maintenance plan while strictly adhering to budget constraints. It operates by selectively exchanging complete annual plans between parent solutions. Specifically, the process consists of two consecutive phases: 1. Probabilistically selecting a reference year; 2. Exchanging the corresponding annual maintenance plans between parent solutions.

[0037] Specifically, such as Figure 2 As shown, from the selected parent individuals, two individuals are selected with a certain probability to serve as parent A and parent B, and a certain year is randomly selected. As a reference year, the first year in parent generation A. Annual maintenance plan (i.e., all assets) (the decision variable) and the parent B's first generation The annual plan is exchanged to generate offspring A' and offspring B'.

[0038] This ensures that offspring inherit the specific year budget allocations from their feasible parents, thereby inherently satisfying annual financial constraints.

[0039] In some feasible embodiments, step S3 specifically includes: Reasonable asset partitioning Then, the granularity of the blocks is dynamically adjusted to appropriately divide them into sub-blocks. This continues until the desired outcome is achieved, forming a multi-layered nested structure. The specific method is as follows: Identify infeasible annual plans. If all annual plans are feasible, randomly select one year for subsequent operations; if multiple infeasible years exist, select the earliest infeasible year in chronological order for processing.

[0040] like Figure 3 As shown, based on geographical location, project affiliation, or asset characteristics, the data for that year is categorized... The assets are divided into individual asset blocks ( The number of assets in each block is ( In particular, the partitioning of assets is not necessarily random or equal; it can be based on geographical location, project affiliation, or asset characteristics.

[0041] For assets that are too large (e.g., 100000) blocks It can be further divided into p sub-blocks. ( The number of assets within the sub-block satisfies This process can be repeated recursively until the final sub-block is of a suitable size, forming a multi-level nested structure.

[0042] In some feasible embodiments, step S4 specifically includes: Recursive logic is used to ensure structured solution space exploration: Hierarchical Operations Planning Mutation: Before executing block crossover, an operations planning mutation operation is performed on the parent block. This involves randomly selecting a certain percentage of assets within the block and forcibly changing their MRR (Mean Restoration Rate) actions. This mutation occurs at each nested level (starting from the top-level block). To its child block It is implemented from top to bottom.

[0043] The interaction between the MRR mutation and the next block crossover operation is determined by the nesting depth and follows a recursive logic: First-level (first-level nesting): If a block Instead of being divided into sub-blocks, it first undergoes MRR mutation. Then, the mutated block continues with the block crossover phase and other subsequent operations.

[0044] Level 2 (double nested level): If the block Divided into sub-blocks ( ), then the parent block First, an MRR mutation occurs. Then, in its sub-blocks... Perform block crossover and other subsequent operations between them.

[0045] Level 3 and above (recursive nesting): This pattern continues recursively. For example, if the child blocks... Further divided into secondary sub-blocks The process is as follows: Parent block MRR mutation occurs; In his son Upper execution block crossover; Then each sub-block Experiencing one's own MRR mutation; Finally, in the second-level sub-block Perform block crossover on top.

[0046] This layered application of variation and crossover is consistently applied to any deeper levels of nesting (e.g., fourth, fifth), ensuring a structured and scalable search strategy.

[0047] like Figure 4 As shown, from each sub-block In this process, groups with medium rankings are selected based on Improved Potential Value (IPV). Assets within these selected sub-blocks are then designated for exchange. These sub-blocks are randomly shuffled and conceptually arranged in a ring, paired sequentially, with the last sub-block paired with the first. Finally, these paired sub-blocks exchange their operational decisions.

[0048] Inter-block maintenance strategy exchange: This occurs between child blocks under the same parent block. Specifically: Calculate the Improvement Potential Value (IPV) for each asset sub-block: in, It is the number of assets in sub-block s. Assets The optimal state that can be achieved. The lower the IPV value (the more negative it is), the greater the potential for improvement.

[0049] Select sub-blocks with a medium IPV ranking for asset swapping to balance exploration and exploitation and avoid premature convergence. Randomly pair the selected sub-blocks and swap the assets specified within them and their corresponding MRR actions.

[0050] In some feasible embodiments, step S5 specifically includes: S5.1, Adaptive capture mechanism; The number K of sub-blocks to be optimized is dynamically calculated. First, based on the sub-block size... Assign it a base selection probability Smaller sub-blocks have a higher probability of being optimized, in order to encourage rapid improvements to sub-blocks that are easy to optimize.

[0051] in, and These are the lower and upper limits of the capture rate.

[0052] The total number of sub-blocks K to be optimized is obtained by taking this... The upper limit of the sum of probabilities of each sub-block is determined.

[0053] like Figure 5 As shown, a "Worst-First" strategy is adopted, which sorts all infeasible sub-blocks according to IPV and selects the top K sub-blocks with the lowest IPV (i.e., the greatest improvement potential) as the capture targets. .

[0054] S5.2 Budget reallocation and LP solution.

[0055] Reallocate the budget for the captured sub-blocks. First, from the total annual budget. Subtract all constraints that are already satisfied from the middle. Budget for each sub-block Then, the remaining budget is allocated to the captured sub-blocks according to certain rules (weighted distribution, equal distribution, or biased allocation based on asset nature). to obtain its new budget : For each captured sub-block, construct and solve the following block-level linear programming model to maximize its local service level: The constraints are: in, The number of assets in the s sub-block participating in the optimization. The service level of the s sub-blocks participating in the optimization in year t. To optimize the state of asset i in sub-block s for year t. To optimize the operation and maintenance costs of asset i corresponding to sub-block s, The budget for the s sub-blocks participating in the optimization. To maintain a tolerance level, actual expenditures are allowed to be within... Fluctuations within the range.

[0056] This operation simplifies complex nonlinear integer programming problems locally into efficient solvable linear programming problems, greatly improving the algorithm's local search efficiency and ability to fix infeasible solutions.

[0057] Based on the above method, the present invention also provides relevant data verification, the results of which are referred to Figure 6 .

[0058] The physical scene is specifically defined as follows: Planning target: the road network of a certain city, which includes 1 million urban and rural roads.

[0059] Initial state: The current average service level of the network (measured by the Road Quality Index, PQI) is 5.86.

[0060] Engineering measures options: Strategy 0: No action taken (performance decays according to the following decay model).

[0061] Strategy 1: Minor repair (PQI improvement of 0.8); Strategy 2: Overhaul (PQI increase by 3); Strategy 3: Rebuild (PQI increase by 10); Project constraints: The annual project budget is 3 billion yuan (with a tolerance of ±20%), and the total budget over 10 years shall not exceed 30 billion yuan.

[0062] Convergence and Efficiency: The algorithm of this invention exhibits excellent convergence performance in this network of millions of assets. After approximately 20 iterations, the objective function (avgLOS) stably converges to the optimal value of 6.19. The entire optimization process takes approximately 80 hours on conventional computing hardware, verifying the algorithm's ability to handle ultra-large-scale problems.

[0063] Feasibility and Quality: After 15 generations of iteration, the constraint penalty term of the population drops to zero, indicating that it can stably generate fully feasible solutions that satisfy 100% of the annual budget constraints. The service level (LOS) and annual expenditure corresponding to the final output solution set both exhibit a stable normal distribution, proving that the planning results are not only optimal but also of highly concentrated and reliable quality.

[0064] Performance Improvement: The final maintenance plan successfully improved the network's average service level from the initial 5.86 to 6.19 over the 10-year planning period, achieving significant optimization of network performance under strict budget constraints and achieving satisfactory maintenance results.

[0065] Technical advantages: Compared with existing technologies, traditional genetic algorithms cannot converge at this scale due to oscillation resolution. This invention successfully solves this computational bottleneck, and its stability and efficiency have been fully verified, establishing its leading position in the field of million-level asset network optimization.

[0066] A large-scale network operation and maintenance system based on nested block genetics includes: The physical modeling module is used to perform step S1; The annual plan exchange cross module is used to execute step S2; The nested module is used to execute step S3; The sub-block processing module is used to execute step S4; The adaptive optimization module is used to execute step S5.

[0067] The content of the above method embodiments is applicable to this system embodiment. The specific functions implemented in this system embodiment are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those achieved in the above method embodiments.

[0068] A large-scale network operation and maintenance device based on nested block genetics: At least one processor; At least one memory for storing at least one program; When the at least one program is executed by the at least one processor, the at least one processor implements a large-scale network operation and maintenance method based on nested block genetics as described above.

[0069] The content of the above method embodiments is applicable to the device embodiments. The specific functions implemented by the device embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0070] A storage medium storing processor-executable instructions, which, when executed by a processor, are used to implement a nested block genetics-based ultra-large-scale network operation and maintenance method as described above.

[0071] The content of the above method embodiments is applicable to this storage medium embodiment. The specific functions implemented in this storage medium embodiment are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those achieved in the above method embodiments.

[0072] The above is a detailed description of the preferred embodiments of the present invention. However, the present invention is not limited to the embodiments described. Those skilled in the art can make various equivalent modifications or substitutions without departing from the spirit of the present invention. All such equivalent modifications or substitutions are included within the scope defined by the claims of this application.

Claims

1. A method for operating and maintaining ultra-large-scale networks based on nested block genetics, characterized in that, Includes the following steps: Acquire asset network data and build an operation and maintenance planning model; Based on the asset network data, a reference year is selected and the corresponding annual maintenance plans are exchanged between the parent generations. For unfeasible annual maintenance plans, all assets within that year are divided into a multi-level nested asset block structure. Within the multi-level nested asset block structure, operational and maintenance variations are performed from top to bottom, and block crossover operations are performed between child blocks under the same parent block. For the last sub-block that does not meet the constraints, perform local optimization operations.

2. The method for operating and maintaining ultra-large-scale networks based on nested block genetics as described in claim 1, characterized in that, The operation and maintenance planning model includes decision variables, objective function, and constraints, wherein: The objective function is to maximize the time-averaged service level of the network during the planning period, and the formula is expressed as follows: in, For the planning period, Total assets For assets The weight, For assets In the Year-end performance status.

3. The method for operating and maintaining ultra-large-scale networks based on nested block genetics as described in claim 2, characterized in that, The step of dividing all assets within a year into a multi-level nested asset block structure for unfeasible annual maintenance plans specifically includes: Identify annual plans that do not meet constraints; According to the preset rules, all assets within the selected annual plan are divided into several primary asset blocks; First-level asset blocks with an asset quantity exceeding a preset threshold are recursively divided into next-level sub-blocks until their asset quantity is less than the preset threshold.

4. The method for operating and maintaining ultra-large-scale networks based on nested block genetics according to claim 3, characterized in that, Also includes: If there are multiple annual plans that do not meet the constraints, the earliest annual plan in chronological order will be selected for processing. If all annual plans meet the constraints, then one annual plan will be randomly selected for processing.

5. The method for operating and maintaining ultra-large-scale networks based on nested block genetics according to claim 1, characterized in that, The step of performing local optimization operations on the last sub-block that does not satisfy the constraints specifically includes: Calculate the basic selection probability of each sub-block based on its size, and determine the sub-blocks to be optimized; Calculate the improvement potential value of the sub-block to be optimized, and construct the set of sub-blocks to be finally optimized; Subtract the budgets of all sub-blocks that have met the constraints from the total annual budget, and allocate the remaining budget to the set of sub-blocks to be finally optimized according to certain rules to obtain a new budget; For each sub-block in the set of sub-blocks to be finally optimized, with the goal of maximizing its local service level, a block-level linear programming model is constructed and solved under the constraint of the new budget.

6. The method for operating and maintaining ultra-large-scale networks based on nested block genetics as described in claim 5, characterized in that, The formula for calculating the total number of sub-blocks to be optimized is as follows: in, This represents the total number of child blocks contained in the parent block. Let be the basic selection probability of sub-block s. Let be the asset size of sub-block s. and These are the preset lower and upper probability limits.

7. The method for operating and maintaining ultra-large-scale networks based on nested block genetics as described in claim 5, characterized in that, The block-level linear programming model is represented as follows: The constraints are: in, The number of assets in the s sub-block participating in the optimization. The service level of the s sub-blocks participating in the optimization in year t. To optimize the state of asset i in sub-block s for year t. To optimize the operation and maintenance costs of asset i corresponding to sub-block s, The budget for the s sub-blocks participating in the optimization. To maintain a tolerance level, actual expenditures are allowed to be within... Fluctuations within the range.

8. A large-scale network operation and maintenance system based on nested block genetics, characterized in that, include: The physical modeling module is used to acquire asset network data and build operation and maintenance planning models; The annual plan exchange cross module, based on the asset network data, selects a reference year and exchanges the corresponding annual maintenance plans between parent generations; The nested block module is used to divide all assets within a year into a multi-level nested asset block structure for unfeasible annual maintenance plans. The sub-block processing module is used to perform MRR mutation from top to bottom within the multi-level nested asset block structure, and to perform block crossover operations between sub-blocks under the same parent block. The adaptive optimization module is used to perform local optimization operations on the last sub-block that does not meet the budget constraints.

9. A large-scale network operation and maintenance device based on nested block genetics, characterized in that, include: At least one processor; At least one memory for storing at least one program; When the at least one program is executed by the at least one processor, the at least one processor implements a large-scale network operation and maintenance method based on nested block genetics as described in any one of claims 1-7.

Images