Reservoir resettlement area site selection method and system

Abstract

The application relates to the technical field of digital site selection planning, and particularly provides a reservoir resettlement area site selection method and system. The method comprises the following steps: constructing a reservoir resettlement area planning model under multi-target optimization, generating an initial site selection planning scheme; performing topological representation on the initial site selection planning scheme to obtain a current resident gathering point-resettlement area distribution matrix, and constructing a solution space; estimating an initial resettlement area quantity; realizing global search by using a simulated annealing algorithm, avoiding local optimization by probabilistically accepting a poor solution, combining the current resident gathering point-resettlement area distribution matrix and an adaptive multi-step neighborhood generation algorithm meeting the resettlement area quantity constraint, iteratively optimizing and adaptively adjusting a search strategy under the feasibility constraint, and converging to obtain an optimal resettlement area site selection scheme. The application organically combines a policy gradient method in reinforcement learning and a simulated annealing framework, and solves the problem of insufficient intelligence of reservoir resettlement area site selection.

Description

Technical Field

[0001] This application relates to the field of digital site selection and planning technology, and in particular to a method and system for site selection of reservoir resettlement areas. Background Technology

[0002] The site selection and allocation of resettlement areas for reservoir migrants is a complex systems engineering project involving social equity, engineering feasibility, and cost optimization. The rationality of the site selection directly affects the quality of life of the migrants, the stability of the regional ecological environment, and the smooth progress of the project. In actual engineering practice, there are numerous alternative resettlement areas with indistinct differences in ecological environment conditions and natural resource endowments. In such scenarios, traditional methods for selecting resettlement areas for reservoir migrants face many intractable technical bottlenecks, resulting in low efficiency and insufficient rationality in site selection and allocation.

[0003] In existing technologies, the site selection of reservoir resettlement areas often relies on manual experience combined with simple mathematical models, resulting in a low level of intelligence. The site selection process lacks a systematic global search mechanism, making it difficult to select the optimal solution from candidate areas with indistinct differences. At the same time, traditional site selection methods often use fixed neighborhood generation, which leads to blind search direction and low search efficiency, further exacerbating the difficulty of site selection and allocation. This makes it difficult to meet the actual engineering needs of reservoir resettlement area site selection, and the site selection process is prone to getting stuck in local optima, making it difficult to find a globally optimal solution that balances the minimum number of resettlement areas with the lowest relocation cost. Summary of the Invention

[0004] To address the aforementioned issues, this application provides a site selection method for reservoir resettlement areas. It employs simulated annealing as the outer framework and an adaptive multi-step neighborhood generation algorithm as the inner neighborhood generation mechanism. This method organically combines the policy gradient method in reinforcement learning with the simulated annealing framework, solving the problem of insufficient intelligence in the site selection of reservoir resettlement areas and overcoming the limitation that the search direction and termination conditions are difficult to determine during the site selection process.

[0005] To achieve the objectives of this application, the following technical solution is provided: In the first aspect, this application provides a method for site selection of reservoir resettlement areas, including: constructing a reservoir resettlement area planning model under multi-objective optimization based on spatial data and basic economic and social data corresponding to the target research area, as well as the requirements of reservoir resettlement areas, and generating an initial site selection planning scheme; The initial site selection planning scheme is represented topologically to obtain the existing residential gathering point-resettlement area allocation matrix based on 0-1 matrix form, and the solution space is constructed; wherein, each element in the existing residential gathering point-resettlement area allocation matrix represents the allocation relationship between residential points and resettlement areas, presenting the mapping from existing residential gathering points to resettlement areas; The initial number of proposed resettlement areas was estimated. And randomly designate within the target study area One alternative resettlement area; Based on the solution space, a simulated annealing algorithm is used to perform a global search, and inferior solutions are accepted with probability to avoid local optima. This is combined with the existing residential cluster-resettlement area allocation matrix and the requirement for a certain number of resettlement areas. The constrained adaptive multi-step neighborhood generation algorithm iteratively optimizes and adaptively adjusts the search strategy under feasibility constraints, and converges to obtain the optimal resettlement area location scheme.

[0006] A further improvement of this application is that the planning model for the reservoir resettlement area includes a site selection objective function and site selection constraints; the site selection objective function includes a first objective function and a second objective function, wherein the first objective function is to minimize the sum of weighted distances from all existing residential gathering points to the proposed resettlement area location, and the second objective function is to minimize the number of proposed resettlement areas; The address selection objective function is specifically as follows: ; ; ; In the formula, This indicates the number of alternative resettlement areas. This indicates the number of current residential gathering points. This represents the population size coefficient of immigrants relocated from existing residential areas. The weighting of resettlement preference is determined by the weighting of centralized resettlement. Indicates the current residential gathering points Assigned to the designated resettlement area The plan; among which, when The time indicates that the first The immigrants from the existing residential areas were allocated to the first Resettlement will be carried out in the reservoir resettlement area. At that time, it was indicated that the first The immigrants from the existing residential areas were allocated to the first A reservoir resettlement area for displaced people; For state description, Indicates except All other allocations besides those; Indicates the current residential gathering points to the proposed resettlement area Distance coefficient, Indicates the current residential gathering points to the proposed resettlement area The traffic cost coefficient; The site selection constraints include comprehensive coverage constraints, flood control and security constraints, spatial capacity constraints, and distance limitation constraints; wherein, the comprehensive coverage constraint specifically refers to: ; ; ; In the formula, This indicates the number of alternative resettlement areas. This indicates the current number of residential settlements; Indicates the current residential gathering points Assigned to the designated resettlement area The plan; Indicates the number of proposed resettlement areas; Indicates the determination of alternative resettlement areas Whether to include it as a variable in the proposed resettlement area, when considering alternative resettlement areas When as a proposed resettlement area =1, otherwise =0; The flood control and security constraints are specifically as follows: ; ; ; in: Resettlement area The elevation corresponding to the flood control standard is given by n, where n represents the return period (in years) corresponding to the flood control standard. This represents the elevation corresponding to a flood defense standard with a return period of n (years). This indicates that the planned relocation and resettlement will be carried out in the alternative resettlement area. The number of immigrants, This represents the population size coefficient of immigrants relocated from existing residential areas. The weighting of resettlement preference is determined by the weighting of centralized resettlement. Indicates the current residential gathering points Assigned to the designated resettlement area The plan This indicates the current number of residential settlements; The spatial capacity constraint refers to the number of immigrants allocated to a candidate resettlement area not exceeding the capacity of the candidate resettlement area and also satisfying the area constraint of the candidate resettlement area. The distance restriction constraint is specifically as follows: ; ; ; in, This refers to mandatory distance restrictions determined based on the actual circumstances of the administrative region. This indicates consideration of the current residential gathering points. The weighted distance limit value is derived after considering topography, economic factors, and social factors. Indicates the current residential gathering points The relocation economic parameters generated by the topography and landforms Indicates the current residential gathering points i The economic parameters of relocation generated by the economy, Indicates the current residential gathering points The relocation economic parameters generated by social factors, Indicates the current residential gathering points to the proposed resettlement area Distance coefficient, This indicates the number of alternative resettlement areas. This indicates the number of current residential gathering points. Indicates the determination of alternative resettlement areas Whether to be considered as a variable for the resettlement area Indicates the current residential gathering points Assigned to the designated resettlement area The plan This represents an economic coefficient that measures the costs, primarily relocation costs, incurred due to distance.

[0007] A further improvement in this application is that the step of performing a topological representation of the initial site selection planning scheme to obtain an allocation matrix of existing residential gathering points and resettlement areas based on a 0-1 matrix, and constructing the solution space, includes: assigning each As an element, it manifests in the reservoir resettlement area. OK 0-1 matrix of columns ,in, Each row contains exactly one 1; based on the 0-1 matrix Define the solution space The solution space Specifically: ; The 0-1 matrix is ​​determined by considering the fixed number of existing residential gathering points and the fixed number of planned resettlement areas in the current stage. At least one row of 1s is randomly moved, and the number of columns whose cumulative sum is not zero remains constant. ; For the solution space For each solution X, the corresponding column and vector are... Each component And based on columns and vectors, calculate the set of zero columns. and the number of zero columns This transforms the activation status of the resettlement area into quantifiable matrix features.

[0008] A further improvement in this application is that, in the reservoir resettlement area planning model, the process of generating neighborhood solutions based on the current solution must satisfy the following condition: There is at least one row , making =1, becomes =1, and j≠k; Number of zero columns in neighborhood solutions With the current solution Number of zero columns Maintain consistency; Solutions in any two solution spaces can be converted to each other through a finite number of neighborhood operations; In the process of generating neighborhood solutions, for the current fixed-dimensional 0-1 matrix form, the basic operation is defined as: selecting rows. and target column , will be done 1 from column Move to column Specifically: ; Columns and updates: ; ; ; To set feasibility conditions for the current column and s, one of the following conditions, A and B, must be met: ; In the formula, To perform a single-point movement operation on the 0-1 matrix of the first element Line 1 The new value of the column, Represents the 0-1 matrix. Line 1 The position of the column corresponds to the current residential gathering point. The originally allocated proposed resettlement area ; Represents the 0-1 matrix. Line 1 The position of the column corresponds to the current residential gathering point. Proposed resettlement areas for relocation ; This indicates that in a 0-1 matrix, except for and The remaining positions outside; For the first 0-1 matrix before the move operation Line 1 The original values ​​of the column; For the original resettlement area after the relocation operation The corresponding column and the new value, Original resettlement area before relocation operation The corresponding columns and; For the original resettlement area after the relocation operation The corresponding column and the new value, Original resettlement area before relocation operation The corresponding columns and; After the move operation Resettlement areas in any other location The corresponding column and the new value, Pre-operation placement area The columns and; It represents the logical "AND".

[0009] A further improvement in this application is that, based on the solution space, a simulated annealing algorithm is used to achieve a global search, and inferior solutions are accepted with probability to avoid local optima. Simultaneously, combined with the existing resident gathering point-resettlement area allocation matrix and an adaptive multi-step neighborhood generation algorithm that satisfies the constraint on the number of resettlement areas, the search strategy is iteratively optimized and adaptively adjusted under feasibility constraints to converge to obtain the optimal resettlement area location scheme. This includes defining an energy function. Calculate the energy change caused by a single point movement. Wherein, the energy function Used to determine the quality of the current solution; parameter initialization is performed on the adaptive multi-step neighborhood generation algorithm and the simulated annealing algorithm, including probability model initialization, operation step distribution initialization, simulated annealing parameter initialization, and solution initialization; rows to be adjusted are selected based on the initialized probability model. and target column Perform a single-point movement operation to generate a neighborhood solution. And in the solution of the neighborhood After feasibility verification, multiple steps are performed according to the initial step distribution to form a composite neighborhood, and the probability model is updated based on the quality of the generated neighborhood solutions. Then, non-first-time neighborhood solutions are obtained based on the updated probability model obtained after the first neighborhood solution. The non-first-time neighborhood solution involves: for the current solution X, determining the row with the maximum row selection probability based on the updated probability model obtained after the first neighborhood solution. And the column corresponding to the uniform distribution of the maximum target column selection probability. After performing a single-point movement operation, a new neighborhood solution is generated. After the feasibility of the new neighborhood solution is checked, the number of steps corresponding to the step number with the highest probability distribution is used as the number of operations to generate a new neighborhood solution. Multi-step operations are performed to form a composite neighborhood. The probability model is updated based on the quality of the generated new neighborhood solution. Then, in the process of algorithm iteration, the non-first neighborhood solution is repeatedly executed. Each time, a new solution is generated based on the probability model updated in the previous round until the termination condition is met, and the optimal solution that meets the constraints is obtained. For each newly generated solution, a simulated annealing acceptance criterion is used to determine whether the new solution can be accepted. If the new solution is accepted, the probability model is updated based on the quality of the new solution until the termination condition is met.

[0010] A further improvement in this application is that, for each newly generated solution, the simulated annealing acceptance criterion is used to determine whether the new solution can be accepted. This includes: determining whether the energy value of the new solution meets the constraint condition of a preset maximum energy value; if the energy value of the new solution meets the constraint condition of the maximum energy value, calculating the energy change value corresponding to the new solution, and using the simulated annealing acceptance criterion to decide whether to accept the new solution based on the energy change value; wherein, the simulated annealing acceptance criterion is: ; In the formula, It is the current temperature and energy change. ;if If the solution is correct, then accept the new solution; otherwise, use probability. Accept the new interpretation.

[0011] A further improvement in this application is that updating the probabilistic model based on the quality of the generated new solution includes: updating the probabilistic model using feedback information from historical searches; specifically including: adaptive updates of row selection probability, column selection probability, step distribution, inverse temperature parameter, and simulated annealing cooling temperature, wherein: The adaptive update of the row selection probability is as follows: ; ; ; In the formula, Indicates line The probability of choosing, Indicates line The quality estimate; The learning rate represents the quality of the row and controls the degree of influence of new information; The inverse temperature parameter selected for row selection is used to control exploration and utilization. The larger the value, the more likely the row with the higher quality estimate will be selected. The smaller the size, the more uniform the distribution; This indicates the reward generated from this move; This indicates assignment; The adaptive update of column selection probabilities is as follows: ; ; in, Indicates the selected row In this case, select column The conditional probability; Indicates the inverse temperature parameter selected in the column; Indicates that will be carried out Move 1 to column The quality estimate; Indicates line i The set of feasible target columns; Indicates the traversal index in the rowizable set; The learning rate indicates the quality of the column; This indicates the reward generated from this move; This indicates assignment; The adaptive update of the step distribution is as follows: ; ; ; in, Indicates the number of steps selected. The probability of performing a neighborhood operation; The adjustment factor selected for the number of steps; Indicates the number of steps to traverse the index; Indicates the maximum allowed number of steps for neighborhood operations; Indicates the number of steps The quality estimate; The learning rate represents the quality of the steps. This represents the total reward for the entire sequence after a single multi-step operation. Indicates the first Energy difference in step operation; Indicates the first The fuzzy annealing temperature during step-by-step operation; The adaptive update of the inverse temperature parameter is as follows: ; In the formula, , They represent the first Second-rate, The inverse temperature parameter generated after the next neighborhood solution is generated. This represents the variance of all row quality estimates. This represents the mean of the quality estimates. ε It is a very small positive number that ensures the denominator is not zero (for example, ε =10 -5 ), Indicates the first The fuzzy annealing temperature during the step operation Indicates the initial temperature; The update of the simulated annealing cooling temperature is as follows: ; In the formula, Indicates the first The temperature of the next time, T k Indicates the first k Temperature at the next iteration Represents the energy function. Represents the best energy in history. This refers to the parameter that controls the decay rate.

[0012] A further improvement of this application is that the termination condition is that the current temperature of the iteration is less than or equal to a preset minimum temperature, or the improvement stagnation condition is reached, or the entropy of the probability model is less than a preset threshold.

[0013] A further improvement in this application is that it also includes: step 5.1, if, when solving for the optimal solution based on the initial number of proposed resettlement areas, a feasible solution that meets the constraints cannot be generated, adjusting... Among the alternative resettlement areas The allocation of proposed resettlement areas is carried out, and the optimal resettlement area site selection scheme is solved again; step 5.2, in adjusting Among the alternative resettlement areas The allocation of proposed resettlement areas is recalculated, and the optimal resettlement area site selection scheme is solved again. If a feasible solution is obtained, adjustments are made step by step to make the optimal resettlement area site selection scheme more feasible. Then, in the adjusted allocation scheme, alternative resettlement areas with assigned values ​​are randomly selected as the new allocation scheme for the proposed resettlement areas. The optimal resettlement area location scheme is solved until no feasible solution can be obtained. In this case, the current number of proposed resettlement areas is determined as the final number of proposed resettlement areas. Step 5.3, in the adjustment Among the alternative resettlement areas If the allocation of the proposed resettlement areas is repeated, and the optimal resettlement area site selection scheme cannot be obtained, then let... And randomly assign values ​​within the target study area. After selecting one alternative resettlement area, the optimal resettlement area location scheme is solved again. If the optimal solution cannot be generated based on the number of proposed resettlement areas after assignment, and if no feasible solution that meets the constraints can be generated, steps 5.1 to 5.3 are executed until the optimal resettlement area location scheme is obtained.

[0014] Secondly, this application provides a reservoir resettlement area site selection system for implementing the aforementioned reservoir resettlement area site selection method, including: The initial scheme generation module is used to construct a multi-objective optimization planning model for reservoir resettlement areas based on the spatial data and socio-economic data corresponding to the target research area, as well as the requirements for reservoir resettlement areas, and to generate initial site selection planning schemes. The solution space construction module is used to perform topological representation of the initial site selection planning scheme, obtain the current residential gathering point-resettlement area allocation matrix based on 0-1 matrix form, and construct the solution space; wherein, each element in the current residential gathering point-resettlement area allocation matrix represents the allocation relationship between residential points and resettlement areas, presenting the mapping from current residential gathering points to resettlement areas; The optimal solution-finding module is used to perform a global search based on the solution space using the simulated annealing algorithm, and to accept inferior solutions with probability to avoid local optima. At the same time, it combines the existing resident gathering point-relocation area allocation matrix and the adaptive multi-step neighborhood generation algorithm that satisfies the constraint of the number of resettlement areas, to iteratively optimize under feasibility constraints and adaptively adjust the search strategy to converge to obtain the optimal resettlement area site selection scheme.

[0015] Compared with the prior art, this application has the following beneficial effects: The reservoir resettlement area site selection method and system provided in this application adopts simulated annealing algorithm as the outer framework and adaptive multi-step neighborhood generation algorithm as the inner neighborhood generation mechanism. It organically combines the policy gradient method in reinforcement learning with the simulated annealing framework, which solves the problem of insufficient intelligence in reservoir resettlement area site selection. It overcomes the limitation that the search direction and termination conditions are difficult to determine in the site selection process. Furthermore, it adopts temperature-aware adaptive learning and adaptive adjustment based on reward variance, realizing the transformation of reservoir resettlement area site selection planning from blind search to intelligent search with memory. While reducing the need for manual parameter tuning, it obtains higher quality solutions than traditional simulated annealing under the same computing resources, and reduces the computation time compared with traditional simulated annealing under the same solution quality requirements. Attached Figure Description

[0016] The accompanying drawings are provided to further understand this application and form part of the specification. They are used together with the embodiments of this application to explain this application and do not constitute a limitation thereof. Figure 1A schematic diagram of one possible process for the site selection method of reservoir resettlement area provided in the embodiments of this application; Figure 2 A schematic diagram of one possible process for the site selection method of reservoir resettlement area provided in the embodiments of this application; Figure 3 This is a schematic diagram illustrating the application process of the adaptive neighborhood generation-intelligent algorithm provided in an embodiment of this application; Figure 4 A schematic diagram illustrating a process for updating the probability model and algorithm parameters provided in an embodiment of this application; Figure 5 A schematic diagram of a modular structure of a reservoir resettlement area site selection system provided in this application embodiment; Figure 6 This is a schematic diagram showing the current distribution of residential gathering points provided in the embodiments of this application; Figure 7 A schematic diagram showing the distribution of alternative reservoir resettlement areas provided in this application embodiment; Figure 8 A schematic diagram showing the location of the reservoir resettlement area and the relocation and resettlement plan provided in the embodiments of this application. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions 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, 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.

[0018] 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 technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature; in the description of this application, unless otherwise stated, "multiple" means two or more.

[0019] The construction of water conservancy projects inevitably leads to large-scale inundation by reservoirs, resulting in the resettlement of reservoir migrants. Resettlement areas, as places for the lives and production of these migrants after relocation, directly impact their quality of life and working conditions. The proper resettlement of these migrants is crucial to the project's construction and benefits, and even to production development and social stability. In practical engineering, there are numerous alternative resettlement areas with indistinct ecological conditions and natural resource endowments. In such scenarios, traditional methods for selecting resettlement areas for reservoir migrants face many intractable technical bottlenecks, leading to low efficiency and insufficient rationality in site selection and settlement allocation.

[0020] Commonly used facility location models, such as set-coverage models, maximum-coverage models, and p-median models, primarily employ graph-based mathematical programming models, typically using the covering method. Their core idea is to select target objects from network nodes that satisfy the optimal objective function under certain constraints. The objective functions of these general methods generally include cost, risk, and fairness objectives. The cost objective is to minimize the transportation cost for the entire population at the proposed locations; the risk objective is to minimize the disaster risk at all proposed locations; and the fairness objective is to achieve the most balanced supply-demand ratio (actual population / capacity) for each facility space. However, when directly applied to the planning of reservoir resettlement areas, traditional models, often based on single-objective or fixed-facility optimization frameworks, struggle to achieve simultaneous collaborative optimization. They cannot reasonably transform multi-objective problems into stable and solvable optimization paths, easily leading to allocation imbalances and infeasible solutions when candidate areas have similar conditions. Furthermore, traditional models rely on fixed constraints and static solution methods, resulting in a highly arbitrary search direction that easily leads to local optima. Additionally, algorithm parameters and neighborhood structures are often manually set, lacking sufficient intelligence. Therefore, in situations where there are no significant differences in ecological and natural resources, site selection and allocation are difficult, and there is a lack of clear objectives, basis, and methods, which is not conducive to progress. At the same time, the process of generating resettlement plans is not intelligent enough, making it difficult to efficiently, accurately, and comprehensively determine site selection plans. This often leads to increased construction and operation management costs, fragmentation of existing residential areas, and is not conducive to long-term stability.

[0021] To address the aforementioned technical problems, this application proposes the following technical solutions and corresponding embodiments.

[0022] The following is combined Figures 1 to 8 The embodiments shown illustrate the technical solutions of this application: Example 1 This application provides a method for selecting a site for a reservoir resettlement area, referring to... Figure 1 As shown, the process includes the following steps S101 to S104: In this embodiment of the application, the total distance of the relocation of immigrants is taken as a factor of feasibility and economy. First, the relocation and resettlement routes are obtained and the distance items are calculated. Then, the site selection model is used to solve the problem and make a comparison of the actual resettlement sites. Finally, the location of the reservoir immigrant resettlement area is determined. After analyzing and verifying the results, the results are presented.

[0023] Step S101: Based on the spatial data and basic economic and social information corresponding to the target research area, as well as the requirements of the reservoir resettlement area, construct a multi-objective optimization planning model for the reservoir resettlement area and generate an initial site selection planning scheme.

[0024] In the embodiments of this application, reference is made to Figure 2 As shown, the number of candidate resettlement areas is first determined based on the spatial data and socio-economic information corresponding to the target study area, as well as the requirements for reservoir resettlement areas. The spatial data and socio-economic information corresponding to the target study area include basic administrative data and an overview of the reservoir project. Basic administrative data includes, but is not limited to, regional area, topography, land use planning, population distribution, economic data, social data, and natural data. The reservoir project overview includes, but is not limited to, the scope of impact, the inundation area, the project scale, flood control standards, resettlement intentions, geological disaster safety, and ecological carrying capacity. Candidate resettlement areas refer to all locations within the land currently permitted for use in the administrative area's planning that can be used for reservoir resettlement.

[0025] In this application embodiment, the requirements for reservoir resettlement areas refer to the comprehensive safety, compliance, livability, and sustainability requirements for centralized resettlement areas for reservoir migrants. These requirements are combined with restrictive factors, including but not limited to threats from floods, mandatory land use planning, core ecological protection areas, complex ownership with no hope of coordination, and situations where laws, regulations, and national policies explicitly prohibit or restrict the resettlement areas. The layout principles are determined around the core objective of protecting the rights and interests of migrants and ensuring the long-term stability of the resettlement areas.

[0026] In this embodiment, spatial data and socio-economic information within the target study area are first identified to obtain basic information on residential settlements (the number of existing residential settlements). Then, in conjunction with the requirements for reservoir resettlement areas, candidate resettlement sites are screened to determine the number of candidate resettlement areas. Specifically, national economic and social development statistics of the relevant administrative regions are collected, and surveys are conducted on the relocated population, houses and ancillary buildings, land, etc. Areas with large-scale residential settlements and a certain number of houses and ground attachments are identified as existing residential settlements, resulting in a total of [number missing]. For each existing residential settlement, determine the location, population size, and relevant parameters for relocation and resettlement (relocation economic parameters generated by topography, relocation economic parameters generated by economic parameters, relocation economic parameters generated by social factors, and economic coefficients of relocation costs mainly caused by distance).

[0027] It is worth noting that this embodiment, based on the goal of fairness, considers construction costs and long-term management, preserves the original customs and features, and requires that resettled residents not be separated. Therefore, in formulating the resettlement plan for reservoir migrants, this embodiment, after considering the resettlement wishes of the migrants, maintains that the existing residential areas are not split or isolated, and is planned uniformly. In this way, by not splitting the existing residential areas, the plan aims to reduce construction costs and long-term operation and management costs, while simultaneously addressing difficulties in production development and social stability, effectively improving the quality of life and production conditions of the reservoir migrants.

[0028] In this embodiment, the three objectives of cost, risk, and fairness are comprehensively considered. The objective function is determined with the goal of minimizing the sum of weighted distances from all settlements to the proposed resettlement area and minimizing the number of proposed resettlement areas. A multi-objective optimization resettlement area planning model (Reservoir Resettlement Area Planning Model) is established to generate an initial site selection plan. Here, the proposed resettlement area refers to the location selected as the resettlement area in the current proposed plan. The reservoir resettlement area planning model not only provides a basis for solving and optimizing the site selection of reservoir resettlement areas but also provides a reference for real-world planning. As a feasible implementation method, constraints such as comprehensive coverage, flood control and security, spatial capacity, and distance limitations are constructed.

[0029] In this embodiment of the application, the per capita area is first determined for the target research area in accordance with policy regulations, and then the area of ​​the resettlement area (reservoir migrant resettlement area) is calculated based on the number of residents currently resettled in the beach area; specifically: Equation (1); Equation (2); in, This indicates the total area (in mu) of the resettlement area. This indicates the number of people resettled in the project. This indicates the per capita area (in mu) that meets safety construction standards. express q The first of the 10 alternative resettlement areas The area of ​​each resettlement area.

[0030] It should be noted that if the total area of ​​the resettlement area cannot meet the per capita area determined in accordance with the policy provisions for all the people to be resettled in the project, it means that the administrative area cannot meet the conditions for resettlement of immigrants, and this situation is not within the scope of this embodiment.

[0031] In this embodiment, the planning model for reservoir resettlement areas includes a site selection objective function and site selection constraints. These constraints include full coverage constraints, flood control and security constraints, spatial capacity constraints, and distance restrictions. The objective function focuses on two objectives: first, minimizing the total cost of the resettlement process, i.e., minimizing the sum of weighted distances from all existing settlements to the resettlement areas; and second, minimizing long-term management and infrastructure construction costs, i.e., minimizing the number of proposed reservoir resettlement areas. In this embodiment, a "0-1" planning model is introduced, with the following mathematical expression: Equation (3); Equation (4); Equation (5); In the formula, This indicates the number of proposed resettlement areas for reservoir migrants. This represents the population size coefficient of immigrants relocated from existing residential areas. The weighting of resettlement preference is determined by the weighting of centralized resettlement. Indicates the current residential gathering points Assigned to the planned reservoir resettlement area The plan Indicates the current residential gathering points To the proposed reservoir resettlement area Distance coefficient, Indicates the current residential gathering points To the proposed reservoir resettlement area The transportation cost coefficient.

[0032] Based on the above embodiments, this embodiment theoretically has a total of [number] possible solutions that can traverse all possibilities. Species, among which This indicates the number of alternative resettlement areas. This represents the number of existing residential settlements. This embodiment first constructs constraints, and through screening based on factors such as comprehensive coverage, flood control and security, spatial capacity, and distance limitations, derives a scheme that meets the constraints. This confirms that the reservoir resettlement plan is correct, reasonable, and feasible, laying the foundation for subsequent optimization using an objective function-oriented approach and incorporating intelligent algorithms and related methods to find the optimal solution.

[0033] Specifically, before starting the calculations, the number of candidate resettlement areas is initially determined, and feedback and adjustments are made in subsequent processes.

[0034] 1) Comprehensive review Ensure that every existing residential area corresponds to a designated reservoir resettlement area, guaranteeing comprehensive coverage. Simultaneously, prevent residents assigned to centralized resettlement areas from being assigned to unsuitable resettlement sites. The details are as follows: Equation (6); Equation (7); Equation (8); In the formula, This indicates the number of alternative resettlement areas. This indicates the current number of residential settlements; Indicates the current residential gathering points Assigned to the planned reservoir resettlement area The plan; This indicates the number of proposed resettlement areas for reservoir migrants; Indicates the determination of alternative resettlement areas Whether to include it as a variable in the resettlement area, when considering alternative resettlement areas When it was used as a resettlement area =1, otherwise =0.

[0035] 2) Flood control and security constraints Based on the classification of protection levels for rural protection zones in the "Flood Control Standard" (GB 50201-2014), and considering the number of immigrants in each resettlement area under the reservoir resettlement plan, the flood control standards are determined and flood control security constraints are established as follows: Equation (9); Equation (10); Equation (11); in: Resettlement area The elevation corresponding to the flood control standard is given by n, where n represents the return period (in years) corresponding to the flood control standard. This represents the elevation corresponding to a flood defense standard with a return period of n (years). This indicates that the planned relocation and resettlement will be carried out in the alternative resettlement area. The number of immigrants, This represents the population size coefficient of immigrants relocated from existing residential areas. The weighting of resettlement preference is determined by the weighting of centralized resettlement. Indicates the current residential gathering points Assigned to the planned reservoir resettlement area The plan This indicates the current number of residential settlements.

[0036] 3) Spatial accommodation constraints Assigned to alternative resettlement areas The number of immigrants shall not exceed the number of those in the alternative resettlement areas. The capacity also meets the area constraints of the alternative resettlement areas. This is stated as follows: Equation (12); Equation (13); Equation (14); Equation (15); in: This indicates that the planned relocation and resettlement will be carried out in the alternative resettlement area. The number of immigrants, Indicates alternative resettlement areas j Environmental capacity limitations Indicates agricultural resettlement. Indicates other composite placement methods, Indicates alternative resettlement areas area, This indicates the per capita area that meets safety construction standards.

[0037] 4) Individual cost constraints (distance constraints); To set distance restrictions and prevent the relocation of certain existing residential areas from being too far away, thus hindering the relocation and reconstruction, the following measures are taken: Equation (16); Equation (17); Equation (18); in, This refers to mandatory distance restrictions determined based on the actual circumstances of the administrative region. This indicates consideration of the current residential gathering points. The weighted distance limit value is derived after considering topography, economic factors, and social factors. Indicates the current residential gathering points The relocation economic parameters generated by the topography and landforms Indicates the current residential gathering points i The relocation economic parameters generated by the economy, Indicates the current residential gathering points The relocation economic parameters generated by social factors, Indicates the current residential gathering points To the proposed reservoir resettlement area Distance coefficient, This indicates the number of alternative resettlement areas. This indicates the number of current residential gathering points. Indicates the determination of alternative resettlement areas Whether to be considered as a variable for the resettlement area Indicates the current residential gathering points Assigned to the planned reservoir resettlement area The plan It represents an economic coefficient that measures the costs, primarily related to relocation, resulting from the influence of distance, natural geography, and socio-economic conditions.

[0038] In this way, policy documents, standards, and laws and regulations related to reservoir resettlement are quantified and mathematically expressed to construct a multi-objective optimization site selection model for reservoir resettlement areas. By clearly defining objectives, basis, and methods, the model solves the difficulties in objectives, constraints, and verification in the site selection planning of reservoir resettlement areas. It can not only provide a basis for solving and optimizing this invention, but also provide a reference for planning schemes in reality.

[0039] Step S102: Perform topological representation on the initial site selection planning scheme to obtain the current residential gathering point-resettlement area allocation matrix based on 0-1 matrix form, and construct the solution space; wherein, each element in the current residential gathering point-resettlement area allocation matrix represents the allocation relationship between residential points and resettlement areas, presenting the mapping from current residential gathering points to resettlement areas.

[0040] In this embodiment of the application, the initial site selection planning scheme refers to the aforementioned The current residential gathering points indicated Assigned to the planned reservoir resettlement area The plan. Based on Indicates the current residential gathering points Assigned to the planned reservoir resettlement area The plan involves topologically analyzing the relocation and resettlement schemes, and then dividing each... The elements form a 0-1 matrix (p rows, q columns) representing the site selection and planning scheme for the reservoir resettlement area; the specific description is as follows: Equation (19); The matrix represents the corresponding planning methods, showing the mapping from the current residential gathering points to the reservoir resettlement area, thus demonstrating the plan. A row in the matrix contains exactly one "1", such as the first row. The "1" in the row is in the first position. The column represents the current residential gathering points. Planning to the reservoir resettlement area .in this way, The matrix serves as the basis for finding the optimal solution. Similarly, this form is also used as the output obtained by intelligent algorithms (such as simulated annealing) after finding the optimal solution.

[0041] In this embodiment of the application, the solution space is defined as all matrices that satisfy the following conditions: Equation (20); It is worth noting that, given that the number of residential gathering points will not change, the required number of rows is [number missing]. Each row contains exactly one element with a value of 1, and the rest are 0; and the number of planned resettlement areas in a current stage under the transformation operation is also fixed. In a 0-1 matrix, this is equivalent to: having at least one row with a unique "1" that is randomly shifted; and maintaining a column count where the cumulative sum of the columns is not zero. This ensures that all solutions within the solution space conform to the basic resettlement logic, thus defining legal boundaries for the generation of subsequent neighborhood solutions. In other words, after obtaining the current resident gathering point-resettlement area allocation matrix, the number of resettlement areas is initially determined, and it is clarified that the number of zero columns in the solution space must remain unchanged, thereby transforming the control of the number of resettlement areas into a matrix structure constraint.

[0042] Furthermore, for each solution X∈ Corresponding to a column and a vector ,in .

[0043] Based on this, define the set of zero columns. Number of zero columns ,as follows: Equation (21); Equation (22); Preserving when generating neighborhood solutions Without changing this requirement, we proceed with the combinatorial optimization problem. New solution generation involves continuously and randomly generating new solutions from the neighborhood of the current solution, and determining whether to accept each new solution as the new current solution, until the stopping criterion is met. New solution generation is the exploratory step of the algorithm, defining how to randomly move from the current solution to another candidate solution.

[0044] In this embodiment of the application, the requirement for neighborhood generation is explicitly stated, that is, for the current solution X, a neighborhood solution is generated. The conditions to be met are stated as follows: ① At least one row needs to be moved: There must be at least one row. , making =1, becomes =1, and j≠k; ② Keep the number of zero columns unchanged: ; ③ It has reachability (connectivity): Any two solutions belonging to the solution space can be transformed into each other through a finite number of neighborhood operations such as transformation and allocation adjustment.

[0045] Here, for the problem of finding the optimal solution for a 0-1 matrix, the neighborhood operations that can be selected include, but are not limited to: basic single-point operations, row operations, column operations, block operations, constraint-preserving operations, etc. Here, we take the single-point movement operation as an example to illustrate the process and method of generating a new solution.

[0046] Furthermore, as an optimization solution to the 0-1 programming problem, the intelligent search algorithm can be any one of the following: simulated annealing algorithm, genetic algorithm, particle swarm optimization algorithm, and optimization algorithm. For example, this embodiment selects the simulated annealing algorithm as an example to illustrate its solution process.

[0047] Furthermore, based on the modified solution for controlling the number of resettlement areas, and for the current fixed-dimensional 0-1 matrix form, the basic operation is defined as: selecting rows. ( (The current column is 1) and the target column. , will be done 1 from column Move to column That is, the existing residential area. The proposed resettlement area for the relocation plan is determined by the resettlement area. Change to resettlement area k The mathematical expression is as follows: Equation (23); Columns and updates: Equation (24); Equation (25); Equation (26); Therefore, to set feasibility conditions for the current column and s, one of the following conditions must be met: Equation (27); In the formula, To perform a single-point movement operation on the 0-1 matrix of the first element Line 1 The new value of the column, Represents the 0-1 matrix. Line 1 The position of the column corresponds to the current residential gathering point. The originally allocated proposed resettlement area ; Represents the 0-1 matrix. Line 1 The position of the column corresponds to the current residential gathering point. Proposed resettlement areas for relocation ; This indicates that in a 0-1 matrix, except for and The remaining positions outside; For the first 0-1 matrix before the move operation Line 1 The original values ​​of the column; For the original resettlement area after the relocation operation The corresponding column and the new value, Original resettlement area before relocation operation The corresponding columns and; For the original resettlement area after the relocation operation The corresponding column and the new value, Original resettlement area before relocation operation The corresponding columns and; After the move operation Resettlement areas in any other location The corresponding column and the new value, Pre-operation placement area The columns and; This represents the logical AND. Thus, the requirements for generating neighborhood solutions as described above can be met.

[0048] Thus, by using a topological relocation and resettlement planning scheme, the site selection planning scheme of the reservoir resettlement area is represented as a matrix, with each element in the matrix representing the corresponding planning method. This not only concisely presents the mapping from the existing residential gathering points to the reservoir resettlement area, but also solves the problem of accurately constructing the solution space in the process of finding the optimal solution in the site selection model.

[0049] Step S103: Estimate the initial number of proposed resettlement areas. And randomly designate within the target study area One alternative resettlement area.

[0050] In this embodiment, the objective function is used to satisfy the fairness goal. Based on the requirement of minimizing the number of resettlement areas, a trial-and-error iterative method is adopted to approximate the overall problem as a single-objective programming problem. Subsequent processes utilize intelligent algorithms to adjust the direction, continuously feeding back and adjusting the number of resettlement areas to achieve the goal of minimizing the number of resettlement areas. Specifically, the steps are as follows: ①Preliminary estimate or plan for the approximate number of high beaches to be constructed .

[0051] ②From Randomly selected from among the candidate reservoir resettlement areas A proposed resettlement area is identified, and then a model algorithm based on a single objective function is used to solve the site selection planning scheme for the reservoir resettlement area.

[0052] ③If a feasible solution can be obtained, for Then return to step ② above. If no feasible solution can be obtained, continue to step ④ below. That is, using the existence of a feasible solution as a constraint, gradually reduce the number of resettlement areas until it can no longer be reduced.

[0053] ④ Adjustment Among the candidate reservoir resettlement areas The allocation of the proposed resettlement areas will continue to employ a single-objective function model algorithm to solve for the planning scheme. If a feasible solution can be obtained, then... Return to step ② above until no feasible solution can be found. That is what I wanted.

[0054] ⑤ If a feasible solution cannot be obtained, for Return to step ② above.

[0055] Therefore, by initially estimating the number of proposed resettlement areas and randomly assigning them... Identify candidate resettlement areas and solve a single-objective planning scheme, then gradually adjust... Until the number of resettlement areas can no longer be reduced, the pre-locking of the number of resettlement areas is completed, and the multi-objective optimization problem is approximately transformed into a single-objective programming problem.

[0056] In this embodiment of the application, the model algorithm based on a single objective function is used to solve the site selection planning scheme of the reservoir resettlement area in accordance with the following steps S104.

[0057] In the current If the adaptive neighborhood generation-intelligent algorithm hybrid optimal solution process fails to find a feasible solution, it is fed back to the modified solution process for controlling the number of resettlement areas. := Processing. Until the optimal solution is obtained "just right". This is the minimum value, and the corresponding solution X is the desired value.

[0058] Step S104: Based on the solution space, a simulated annealing algorithm is used to perform a global search, and inferior solutions are accepted with probability to avoid local optima. At the same time, combined with the existing resident gathering point-resettlement area allocation matrix and an adaptive multi-step neighborhood generation algorithm that satisfies the resettlement area quantity constraint, the search strategy is iteratively optimized and adaptively adjusted under the feasibility constraint, and the optimal resettlement area site selection scheme is obtained by convergence.

[0059] In this application embodiment, the goal is to find the optimal solution that minimizes the overall cost of relocation, resettlement, and management in the problem of site selection for reservoir resettlement areas. The application identifies a practical method for quickly and effectively finding satisfactory solutions using intelligent algorithms, such as metaheuristic algorithms. This method provides a general process and approach applicable to solving various optimization problems, maintaining one or more current solutions (or a population of solutions) and generating new solutions through a series of operations (such as perturbation, recombination, and selection). These operations typically involve randomness, thus the algorithm generates a stochastic process. ,in Let the state of generation t be a single solution or a set of solutions. The goal of the algorithm is to make... The corresponding objective function value approaches the optimal value as t increases.

[0060] It should be noted that the total cost includes, but is not limited to, direct compensation and subsidies, construction costs, and relocation costs. This section focuses on factors that cause differences in resettlement plans during the relocation process; other situations and factors are also applicable. The algorithm in this embodiment is also compatible with the determination of the energy function.

[0061] This embodiment employs a metaheuristic algorithm to quickly and effectively find a satisfactory solution, defining the quadruplet population update operator, individual optimization operator, selection operator, and termination criterion as follows: (1) Population update operator: How to conduct global exploration and development. For two parent allocation schemes, the allocation relationship between some existing settlements and the proposed resettlement area is exchanged with a certain probability, or the allocation of several existing settlements in the proposed resettlement area in a single solution is randomly changed.

[0062] (2) Individual optimization operators: How to perform local fine search. Used for deep optimization of a single solution, such as using exchange-based local search: moving a residential area to another resettlement site, exchanging the resettlement site allocation of two residential areas, etc.

[0063] (3) Selection Operator: How to guide the direction of population evolution. The selection operator is used to select individuals from the current population to enter the next generation. It is usually based on fitness (objective function value), such as relocation cost function: including distance cost, construction cost, resettlement quality, infrastructure, environmental suitability, etc.

[0064] (4) Termination Criteria: How to determine if the algorithm has converged. Determine when to stop the search, and obtain the site selection scheme for the reservoir resettlement area through a finite number of solutions, balancing computational cost and solution quality.

[0065] The process of determining the optimal solution serves two objectives: exploration and utilization. In terms of exploration, factors such as step distribution, novelty, and column distance encourage the algorithm to try new regions. In terms of utilization, factors such as contribution, historical improvement, and column balance guide the algorithm to search in promising directions.

[0066] Therefore, based on the solution space, an adaptive neighborhood generation-intelligent algorithm hybrid approach is used to find the optimal solution. This involves continuously generating neighborhood solutions, i.e., matching and mapping relationships between different existing settlements and the proposed resettlement area. During the process, an intelligent algorithm is used to continuously adjust the direction of solution generation and to judge whether each new solution meets the constraints and its quality, thereby adjusting / adopting / terminating the process. This embodiment uses simulated annealing as an example to illustrate the method steps: each complete iteration (generating candidate solutions → calculating energy differences → accepting based on probability → cooling) corresponds to a state transition in a Markov chain. The entire annealing process constitutes a non-homogeneous (time-varying) Markov chain because the state transition probability changes continuously as the temperature decreases.

[0067] The core of a Markov chain is the state and the state transition probability. Each feasible 0-1 matrix X∈ It's a state in the chain, mapping from the current settlement to the proposed resettlement area plan, i.e., state X, and then transitioning to the next updated state. State transition probability P( ) is generated by probability G( ), acceptance probability A( The two stages together determine this, as described below: ; like If it is infeasible (constraints are not met) or cannot be generated within the maximum number of attempts, then .

[0068] Furthermore, regarding the selection strategy for neighborhood operations: on the one hand, perform an operation feasibility check by substituting the current solution into the model to verify whether it meets the constraints, as part of the solution process; on the other hand, design a probability distribution; to overcome the limitations and misleading effects of uniform distribution, determine how to select specific moves when generating neighborhood solutions from the current solution, and combine simulated annealing algorithm to make the search direction tend towards energy decrease (the quality of the solution is improved, see the energy function defined in S1031 for details), while maintaining a certain exploration capability.

[0069] In this embodiment, the simulated annealing algorithm is used to control the search process using temperature parameters. When using the simulated annealing algorithm, the initial temperature, termination temperature, cooling function, energy function, solution neighborhood, and movement parameters must be provided. The goal of the annealing algorithm is to find the solution that minimizes the energy function. Starting from the initial solution... and control parameters initial temperature value The process begins with repeated iterations: generating neighborhood solutions of the current solution - calculating the energy function difference - accepting or discarding solutions while simultaneously decreasing the temperature. T The value when T The algorithm terminates when the value falls below the termination control parameter, and the current historical best solution is the optimal solution.

[0070] As a feasible implementation method, this embodiment embeds the adaptive neighborhood solution generation algorithm into the simulated annealing algorithm framework to derive the optimal solution implementation method path. (Refer to...) Figure 3 The specific implementation method path is described as follows: Step S1031: Define the energy function Calculate the energy change caused by a single point movement. ; Specifically, the energy function is defined as follows: Equation (28); Through energy function The value of is used to judge the quality of the current solution; the lower the value, the closer the corresponding solution is to the optimal solution.

[0071] Based on this, we can explain the local approximation of energy change for row... Currently, 1 is listed. Consider moving to column k; the energy change is expressed as follows: Equation (29); It is used to measure the quality of a solution and to determine whether to accept a new solution based on the energy difference between the new and old solutions.

[0072] Here, to avoid solutions with excessive energy consumption, a total energy limit is set based on the actual situation of the relocation and resettlement area and the affected zone. .

[0073] Step S1032: Initialize the parameters and models of the adaptive neighborhood generation algorithm and the simulated annealing algorithm; In this embodiment of the application, the specific process of initializing the parameters of the probability model and algorithm can be implemented through the following steps S10321 to S10324: Step S10321, in the initial (current) state Below, single-point movement operation selects a row. and target column The probability is initialized to a uniform distribution, and the uniformly initialized probability model is expressed as follows: Equation (30); Equation (31); in: This represents the probability of row selection being uniformly distributed. This indicates the probability of selecting a uniform distribution for the target column. Indicates a row The set of all feasible target columns (feasibility is determined by constraints, such as keeping the number of zero columns constant). The number of feasible target columns.

[0074] In this way, an initial probability distribution is established for the operation in each state to guide subsequent searches.

[0075] It should be noted that the probabilistic model is used to generate neighborhood solutions in step S1033 and is updated in step S1035 based on search experience.

[0076] Step S10322: Initialize the distribution of operation steps; The geometric distribution has memorylessness and is suitable for describing the number of steps in a stochastic decision-making process; therefore, the truncated geometric distribution is used as the step distribution.

[0077] Equation (32); in, L This indicates the number of basic operation steps included in a single multi-step generation. This indicates the preset maximum number of steps. l This indicates the specific value that can be taken for the number of steps. ρ Indicates the attenuation coefficient (0 < ρ <1). Truncation ensures the number of steps. L No more than To control computing costs, and by adjusting ρ Controlling the perturbation intensity of the initial search: ρ The closer the distribution is to 1, the flatter it is; the closer it is to 0, the more concentrated it is. L =1; a smaller ρ makes single-step operations dominant, and the search is more local; a larger ρ increases the probability of multi-step operations, and the search is more global.

[0078] Step S10323: Initialize simulated annealing parameters; In the embodiments of this application, it is estimated that at a given temperature T The algorithm accepts the average probability of a solution from a random neighborhood, and from this, it deduces the temperature at which the average acceptance probability reaches the target value, thus setting a reasonable starting point for the annealing process. The approximate formula for the average acceptance probability is as follows: Equation (33); in: Indicates temperature T The average acceptance probability is as follows: M This indicates the number of samples sampled from a random neighborhood. Indicates the first m The energy difference between the new solution and the current solution in the next sampling T This represents the temperature parameter.

[0079] Further, calculate the initial temperature. Assuming that negative energy differences (energy decreases) are always accepted, then the average acceptance probability can be approximated as the average acceptance probability of positive energy differences. Let it equal... χ 0, and assume M + / M is close to 1 (because both positive and negative changes are possible at high temperatures, but positive changes are likely to be slightly more frequent), and taking the logarithm, we get the following: Equation (34); Equation (35); in: This represents the initial acceptance rate of the target (usually set to 0.7~0.9).

[0080] This allows the algorithm to have the desired acceptance rate and more even distribution in the initial stage, thus enabling it to access various regions in the solution space with a higher probability, laying the foundation for global exploration.

[0081] Step S10324: Initialization of the solution; In this embodiment, an initial trial solution is obtained by arbitrarily selecting a planning scheme for a reservoir resettlement area from an existing residential settlement that meets the constraints of the site selection model. Given the termination parameter value .

[0082] In this embodiment, the probability distribution design is equivalent to a distribution model of the solution space, used to guide the generation of new solutions; it is also integrated into the simulated annealing parameters to control the probability of accepting poor solutions, which decreases as the temperature decreases. The initialization settings provide "blank space" and "heat" for optimization, meaning the algorithm will not miss certain regions due to initial preferences, nor will it prematurely fall into local optima due to excessively low temperatures. This ensures that as the simulated annealing temperature decreases in the later stages of the search, the algorithm will not get stuck in a local optimum prematurely due to excessive "conservatism" and only accepting the best solution. Therefore, as iterations proceed, the decreasing temperature and policy updates will gradually guide the chain towards more promising regions, ultimately converging to a high-quality solution.

[0083] Step S1033: In each process of generating a neighborhood solution, perform multiple operations to generate a neighborhood solution; In this embodiment, a neighborhood solution is generated through multiple neighborhood operations. During the first solution calculation, the initialized model is executed directly. Subsequent neighborhood solution calculations first adjust updates related to the probability model, and then the updated model is used for the solution. Specifically, during the first neighborhood solution calculation, the solution is obtained based on the initialized adaptive neighborhood generation algorithm and simulated annealing algorithm model constructed in step S1032; this is implemented through the following steps S10331a to S10331d: Step S10331a: Row and column selection; For a given initial solution According to step S10321 , Determine the row with the highest uniform distribution probability. and for the first Rows determine the maximum target column; columns are selected based on the uniform distribution probability. .

[0084] Step S10331b: Generate a neighborhood solution after the swap; For the current initial solution , will the The line was originally located in the first The element 1 in column k is moved to row k, while the rest of the elements remain unchanged. After row and column selection and swapping operations, the resulting neighborhood solution... Its constituent elements are described as follows: Equation (36); Step S10331c: Feasibility verification; In this application embodiment, on the one hand, it is necessary to meet the topological matrix condition constraints. Taking maintaining zero column number as an example, the feasibility conditions for the current column and s (the column and vector of the current solution) must meet one of the following conditions (i.e., equation (27)): Equation (37); In the formula, Indicates the original resettlement area before the relocation operation. The corresponding column sum, i.e., the original resettlement area before the relocation. The number of settlements allocated; Indicates the original resettlement area before the relocation operation. The corresponding column sum, i.e., the original resettlement area before the relocation. The number of settlements allocated, It represents the logical "AND"; its purpose is to keep the number of zero columns unchanged, that is, to keep the number of resettlement areas unchanged.

[0085] On the other hand, to ensure that the constraints are always satisfied in multi-step operations, after finding the neighborhood solution, the... Substitute the constraints and verification conditions into the location selection model to ensure it passes. If the verification fails, repeat the above process: supplement the row selection probability and the maximum target column selection probability, and then perform the swap to obtain a new neighborhood solution.

[0086] Step S10331d: Multi-step operation; In this embodiment of the application, according to equation (32) in step S10322 above, the step number distribution probability corresponding to the maximum is taken. l The number of operations is used to generate neighborhood solutions, and then the contents of steps S10331a and S10331b above are executed again.

[0087] Thus, multi-step operations ( L >1) It can construct an exponentially expanding composite neighborhood, enabling the algorithm to reach a more distant solution space region in one step, greatly enhancing its ability to escape local optima.

[0088] Step S1034: Determine whether the new solution can be accepted based on the simulated annealing acceptance criterion; In this embodiment, the first step is to determine whether the energy value of the new solution meets the constraint condition of the maximum energy value, as described below: ; If the above conditions are met, the acceptance criterion can be applied. Otherwise, return to step S1033 to regenerate the neighborhood solution.

[0089] Furthermore, the simulated annealing acceptance criterion (Metropolis Monte Carlo criterion) is applied to the metaheuristic algorithm, using the generated neighborhood solutions (new solutions) to determine whether to accept the new solutions based on energy changes. This helps the algorithm escape local optima and achieve global search. Specifically, the simulated annealing acceptance criterion is as follows: Equation (38); In the formula, It is the current temperature and energy change. ;if If the solution is correct, then accept the new solution; otherwise, use probability. Accept the new interpretation.

[0090] Here, the probability of accepting poor solutions decreases as the temperature decreases; exploration is allowed initially, but exploitation is preferred later. This ensures that the algorithm can accept poor solutions at high temperatures for global exploration, while primarily accepting good solutions for local search at low temperatures. The acceptance criterion depends on energy changes. and current temperature Meanwhile, the accepted results will be used to update the probability model in the subsequent step S1035.

[0091] Step S1035: Update the probability model and algorithm parameters according to the operation results to obtain the adaptive adjustment model, and update the generated new solution based on the adaptive adjustment model until the termination condition is met to obtain the optimal solution; In the embodiments of this application, reference is made to Figure 4 As shown, the probability model is updated based on the quality of the newly generated solutions, and feedback information from historical searches (such as the reward for each operation) is used to update the probability model, allowing the algorithm to gradually focus on more effective search behaviors. By adaptively adjusting the search direction, the probability of generating good solutions increases; specifically including: S1035a, Adaptive update of row selection probability: Equation (39); Equation (40); Equation (41); In the formula, Indicates line The probability of choosing, Indicates line The quality estimate (initially set to 0 or a specified prior value); The learning rate represents the quality of the rows and controls the degree of influence of new information (usually a small value, e.g., ...). ); The inverse temperature parameter selected for row selection is used to control exploration and utilization. The larger the value, the more likely the row with the higher quality estimate will be selected. The smaller the size, the more uniform the distribution; This indicates the reward generated from this move; This indicates the assignment of values. In this way, based on historical experience, rows with good performance are given a higher probability of selection, but some exploratory nature is retained, so that the quality estimate smoothly reflects recent returns, thereby gradually adjusting the selection probability.

[0092] As a feasible implementation method, in single-point movement, the adaptive update objective of the row selection distribution can be to increase the probability of selecting rows that historically generated high returns (i.e., significant energy reduction) in the current state. Essentially, it involves adjusting the resettlement plan based on the existing resident resettlement sites that maximize the optimization of the planning scheme. A value function-based method maintains a value estimate for each (state, row) pair. The row selection probability is represented as follows: Equation (42); In the formula, The parameter vector representing the row selection strategy. This represents the scoring function that evaluates the score for row i. The scoring function can be a deep neural network. As an optional approach, it will not be elaborated here.

[0093] S1035b, Adaptive update of column selection probability: Equation (43); Equation (44); in, Indicates the selected row In this case, select column The conditional probability; Indicates the inverse temperature parameter selected in the column; Indicates that will be carried out Move 1 to column The quality estimate (initially set to 0 or a specified prior value); Indicates line i The set of feasible target columns (determined by constraints); Indicates the traversal index in the rowizable set; The learning rate indicates the quality of the column; This indicates the reward generated from this move; This indicates an assignment. Thus, for a given row, priority is given to target columns that have historically yielded high returns. Above, the denominator is a normalization constant, ensuring that the sum of the probabilities of all feasible target columns is 1.

[0094] In summary, moves with greater energy decreases have a higher probability of being selected, but moves with increasing energy also have a certain probability of being selected, which helps to escape local optima. This guides the contraction direction towards the direction of decreasing energy while maintaining randomness to avoid premature convergence.

[0095] S1035c, adaptive update of step distribution; In the embodiments of this application, for each possible step number l (from 1 to ... Maintain a quality estimate This reflects the overall effect of using this number of steps to perform multiple operations.

[0096] Equation (45); Equation (46); Equation (47); in, Indicates the number of steps selected. The probability of performing a neighborhood operation; The adjustment factor selected for the number of steps; Indicates the number of steps to traverse the index; Indicates the maximum allowed number of steps for neighborhood operations; Indicates the number of steps The quality estimate (initially set to 0 or a specified prior value); The learning rate represents the quality of the steps. This represents the total reward for the entire sequence after a single multi-step operation. Indicates the first Energy difference in step operation; Indicates the first The fuzzy annealing temperature during the step operation.

[0097] Therefore, the step size is adaptively adjusted based on the current temperature and the diversity of solutions to balance exploration and exploitation. At high temperatures, the probability model is more evenly distributed, and a larger step size is used to enhance exploration; at low temperatures, the probability model may be concentrated on a few operations, and a smaller step size is used for local fine-tuning, thus avoiding overfitting while maintaining a balance between exploration and exploitation.

[0098] S1035d, adaptive update of inverse temperature parameters; In this embodiment of the application, a small initial β is set to set the initial distribution as a uniform distribution, which is then adaptively adjusted subsequently.

[0099] Equation (48); In the formula, , They represent the first Second-rate, The inverse temperature parameter generated after the next neighborhood solution is generated. This represents the variance of all row quality estimates. This represents the mean of the quality estimates. ε It is a very small positive number that ensures the denominator is not zero (for example, ε =10 -5 ), Indicates the first The fuzzy annealing temperature during the step operation This indicates the initial temperature.

[0100] In this way, it adaptively adjusts based on the search status, increasing when the variance of the quality estimate is large. Increase when temperature decreases This achieves a three-level adaptive process involving quality estimation, selection probability, and inverse temperature parameters.

[0101] S1035e, Updated Simulated Annealing Cooling Temperature: Equation (49); in, Indicates the first The temperature of the next time, T k Indicates the first k Temperature at the next iteration; This represents the cooling coefficient, which can be a constant (typically 0.95) or a temperature-dependent function.

[0102] Here, the cooling schedule specifies how the temperature decreases as iterations proceed, thereby gradually reducing the probability of accepting inferior solutions and shifting the algorithm from exploration to exploitation; temperature The probability of the algorithm accepting inferior solutions is controlled. As iterations proceed, the temperature gradually decreases, and the algorithm shifts from exploration to exploitation. Thus, the initial temperature of simulated annealing and the rules for temperature decrease are set.

[0103] As a feasible implementation method, the cooling coefficient Define an adaptive temperature-dependent function and introduce parameters to control the decay rate. The adaptive temperature adjustment is represented as follows: Equation (50); In the formula, Indicates the first The temperature of the next time, T k Indicates the first k Temperature at the next iteration Represents the energy function. Represents the best energy in history. This refers to the parameter that controls the decay rate.

[0104] Thus, the temperature The descent is defined as relying on the relative difference between the energy of the current solution and the historical best energy. If the current solution is far from being optimal, the temperature is lowered to make the probability distribution sharper, thus rapidly improving the solution; if the energy of the current solution is close to the historical best energy, ≈ Maintaining a higher temperature increases the search speed and avoids getting trapped in local optima; thus, adjustments are made based on the search progress to improve convergence speed and global search capability.

[0105] In this way, the neighborhood generation algorithm adopts an adaptive probability distribution design, transforms the quality estimate into a probability distribution through the Boltzmann distribution, and ensures that the reservoir resettlement area site selection planning achieves a balance between exploration (trying other options) and utilization (selecting the option with the best current estimate) during the solution process, based on the confidence level of the probability model and temperature adaptive step size control, thus realizing strategic search.

[0106] Therefore, based on the adaptive adjustment model obtained in step S1035 above, a non-first-time neighborhood solution is calculated; the content of the non-first-time neighborhood solution specifically includes: Step S10351a: Row and column selection; For the current solution X, based on the steps in equations (39) to (41) and (43) to (44) above, determine the row corresponding to the maximum row selection probability after adaptive update. and for the first The row determines the column corresponding to the uniform distribution of the maximum target column selection probability after adaptive update. .

[0107] Step S10351b: After the exchange, a neighborhood solution is generated; The implementation of this step is consistent with the method of generating neighborhood solutions after the exchange in step S10331b above.

[0108] Step S10351c, Feasibility verification; The implementation of this step is consistent with the feasibility verification method in step S10331c above. That is, it requires feasibility conditions that meet the topological matrix constraints as shown in equation (37). At the same time, after finding the neighborhood solution, the neighborhood solution is substituted into the constraints and verification conditions of the location model to ensure that it can pass.

[0109] Step S10351d, multi-step operation; In this embodiment of the application, according to the above equations (45) to (47), the value corresponding to the operation step distribution probability with the highest value is taken. The number of operations used to generate a neighborhood solution is used to execute steps S10351a and S10351b. Here, a multi-step operation, i.e., one neighborhood generation, consists of L basic operations, which can be represented as follows: Equation (51); In the formula, Indicates an L-step compound operation. This represents the Lth step, a single-point movement operation (feasible case). Thus, the input to each step of the operation is the output of the previous step, and the feasibility of subsequent operations depends entirely on the dependencies of the current state.

[0110] Furthermore, the state transition is described as follows: Equation (52); In the formula, Indicates the first The state of the step; Indicates the first The state of the step, that is, relative to the first step. The state of the previous step; Indicates the first Neighborhood operation operator of step, This represents the operational parameters obtained from the adaptive adjustment model.

[0111] In this way, by performing multiple neighborhood operations consecutively, a new solution far removed from the current solution can be generated, thus enhancing the exploration capability. A single-step operation may only be able to search within a small range, while a multi-step operation can reach a more distant region in the solution space.

[0112] Therefore, based on the newly generated neighborhood solution, the process is repeated from step S1034 to step S1035, obtaining the updated adaptive probability model until the termination condition is met; that is, steps S1034-S1035 are repeated until the termination condition is met. In this embodiment, the termination condition may be reaching the maximum number of iterations or the temperature being below a certain level. Or, if there is no improvement after several consecutive iterations; in addition, the above termination conditions can also be used in combination to balance the quality of the solution during the calculation.

[0113] As a possible implementation method, the termination condition is a temperature condition, that is, the termination occurs when the temperature drops below the minimum temperature: ; In the formula, This represents the current temperature at the k-th iteration of the algorithm. This represents the preset minimum temperature threshold (a very small positive number).

[0114] when When this condition is met, it indicates that the system has "frozen," the probability of accepting any inferior solution is extremely low, and the algorithm has essentially stopped effective global exploration. Continuing to iterate will contribute very little to improving the quality of the solution. This condition ensures the physical completeness of the algorithm's annealing process.

[0115] As a possible implementation method, when the termination condition is an improved stagnation condition, i.e., in the most recent In this iteration, the optimal solution did not show significant improvement, specifically: ; In the formula, Indicates as of the date The historical best energy value found by the next-neighborhood solution generation algorithm This represents the size of the observation window (positive integer, e.g., representing the span of the historical iterations) that reflects the historical iterations. =50), This represents the relative improvement threshold (a very small positive number, for example, =10 -5 ), to determine whether the improvement is significant.

[0116] Therefore, this indicates the speed of algorithm optimization. If, within the observation window, the relative improvement of the optimal solution consistently falls below a small threshold, it means the algorithm has plateaued. This could be due to a global optimum or a deep local optimum from which it is unlikely to escape, and further searching will yield very little benefit. Stopping the algorithm under this condition avoids unnecessary computation in regions that have already converged.

[0117] As a feasible implementation method, when the termination condition is the convergence condition of the probabilistic model, that is, the entropy of the probabilistic model is less than a certain threshold, it means that the probability distribution has become very concentrated, and certain operations are selected almost deterministically, specifically: ; ; In the formula, Representing a probabilistic model Information entropy Entropy represents the probability distribution. This represents the probabilistic model (row and column selection, number of steps, etc.) in the adaptive neighborhood solution algorithm during the k-th iteration. This represents the maximum entropy corresponding to a uniform distribution on the same support set. ), This represents the probability value corresponding to the parameter. The basic unit of calculation representing entropy; Indicates the size of the operation parameter space. This represents the threshold of the probability set (a positive number close to 1, for example...). =0.99).

[0118] Thus, the concentration of the probabilistic model is measured, such as This means that the probability distribution is highly concentrated in a few operations, indicating that the model has moved from the learning phase to the "confidence" phase and has completed its exploration strategy. This ensures that the algorithm terminates promptly when it stops exploring.

[0119] Therefore, by setting reasonable termination conditions, the algorithm can stop at an appropriate time, avoiding infinite running or premature termination. These conditions depend on variables generated during the algorithm's execution, such as temperature. T Optimal energy E best and probability model P By judging whether the algorithm has converged from different perspectives, the robustness of the termination judgment is improved. Finally, based on the statistical decision-making process of probability guarantee, computational budget and convergence diagnosis, the algorithm determines whether to stop the operation and confirm the optimal solution, so that the algorithm has asymptotic global convergence in theory, but in practice it must stop within a finite time.

[0120] It should be noted that if this embodiment cannot generate an optimal solution that meets the constraints in the current module, it indicates that the solution deformation module is ineffective. num r If the value is not correct, you can return to the deformation solution module for further processing. num r = num r +1 is processed, and the process of finding the optimal solution for the current module is repeated.

[0121] The framework of this embodiment is also applicable to a variety of combinatorial optimization problems (scheduling, allocation, etc.), which can be applied simply by defining an energy function and neighborhood operations; and it is easy to integrate other metaheuristic components, supporting distributed and parallel implementations.

[0122] The reservoir resettlement area site selection method provided in this embodiment adopts simulated annealing algorithm as the outer framework and adaptive multi-step neighborhood generation algorithm as the inner neighborhood generation mechanism. It organically combines the policy gradient method in reinforcement learning with the simulated annealing framework, which solves the problem of insufficient intelligence in the site selection of reservoir resettlement areas. It overcomes the limitation that the search direction and termination conditions are difficult to determine in the site selection process. Furthermore, it adopts temperature-aware adaptive learning and adaptive adjustment based on reward variance, realizing the transformation of the site selection planning of reservoir resettlement areas from blind search to intelligent search with memory. While reducing the need for manual parameter tuning, it obtains higher quality solutions than traditional simulated annealing under the same computing resources, and reduces the computing time compared with traditional simulated annealing under the same solution quality requirements.

[0123] Example 2 Based on the above embodiments, this embodiment also provides a reservoir resettlement area site selection system for implementing the above-described reservoir resettlement area site selection method. The reservoir resettlement area site selection system of this embodiment includes: The initial scheme generation module is used to construct a multi-objective optimization planning model for reservoir resettlement areas based on the spatial data and socio-economic data corresponding to the target research area, as well as the requirements for reservoir resettlement areas, and to generate initial site selection planning schemes. The solution space construction module is used to perform topological representation of the initial site selection planning scheme, obtain the current residential gathering point-resettlement area allocation matrix based on 0-1 matrix form, and construct the solution space; wherein, each element in the current residential gathering point-resettlement area allocation matrix represents the allocation relationship between residential points and resettlement areas, presenting the mapping from current residential gathering points to resettlement areas; The optimal solution-finding module is used to perform a global search based on the solution space using the simulated annealing algorithm, and to accept inferior solutions with probability to avoid local optima. At the same time, it combines the existing resident gathering point-relocation area allocation matrix and the adaptive multi-step neighborhood generation algorithm that satisfies the constraint of the number of resettlement areas, to iteratively optimize under feasibility constraints and adaptively adjust the search strategy to converge to obtain the optimal resettlement area site selection scheme.

[0124] Example 3 Based on the foregoing embodiments, this embodiment also provides a site selection system for reservoir resettlement areas, referring to... Figure 5 As shown, the reservoir resettlement area site selection system in this embodiment includes 5 modules, wherein modules one through four are performed sequentially, and there is feedback adjustment between modules four and five.

[0125] The module consists of five modules: Module 1: Identification and judgment of the current spatial status; Module 2: Construction of a multi-objective optimization site selection model for reservoir resettlement areas; Module 3: Solution of the topological mapping matrix between the current status and resettlement; Module 4: Solution transformation for controlling the number of planned resettlement areas; and Module 5: Hybrid solution of adaptive neighborhood generation and intelligent algorithm for finding the optimal solution.

[0126] In Module 1, spatial data and socio-economic information about the study area are input to identify basic information about existing residential settlements. Further, by considering multiple dimensions such as safety, compliance, livability, and sustainability of the reservoir resettlement area, and taking into account limiting factors including but not limited to flood threats, mandatory land use planning, ecological protection core areas, complex ownership structures with no hope of coordination, and situations explicitly prohibited or restricted by laws, regulations, and national policies, the layout principles are determined around the core objective of protecting the rights of migrants and ensuring the long-term stability of the resettlement area, and candidate resettlement sites are selected. Based on the characteristics and requirements of the reservoir resettlement area, basic information and data are compiled and organized. A total of existing residential settlements are identified. pThere are [number] eligible candidate resettlement areas for reservoir migrant resettlement. q indivual.

[0127] In Module Two, the site selection problem is analyzed, and a multi-objective optimization model for resettlement area planning of reservoir migrants is established. This model comprehensively considers three objectives: cost, risk, and equity. The objective function is determined to be guided by minimizing the sum of weighted distances from all settlements to the proposed resettlement area and minimizing the number of resettlement areas. Correspondingly, constraints such as comprehensive coverage, flood control and security, spatial capacity, and distance limitations are constructed. This model not only provides a basis for solving and optimizing this invention but also serves as a reference for real-world scheme planning.

[0128] In Module 3, the relocation and resettlement plans are represented topologically, with the site selection plans for the reservoir resettlement area presented as a matrix. Furthermore, each element in the matrix represents a corresponding planning method, showing the mapping from existing residential areas to the reservoir resettlement area, thus revealing the plans. This serves as the basis for finding the optimal solution.

[0129] In Module 4, on the one hand, the objective function is used to meet the fairness objective; on the other hand, the multi-objective optimization model is simplified based on the requirement of minimizing the number of resettlement areas. The solution process is simplified by using a "trial-and-error" iterative method to determine the number of resettlement areas, which is approximately transformed into a single-objective programming problem. In addition, the direction adjustment of the intelligent algorithm is used to continuously provide feedback and adjust the number of resettlement areas to achieve the goal of minimizing the number of resettlement areas.

[0130] In Module 5, a metaheuristic intelligent algorithm is used as a means to find the optimal solution, and it is creatively combined with an adaptive neighborhood solution generation algorithm. The two are used together. Taking the annealing algorithm as an example to illustrate the application method, the simulated annealing algorithm is used as the outer framework, and the adaptive multi-step neighborhood generation algorithm is used as the inner neighborhood generation mechanism. The temperature parameter of the simulated annealing algorithm controls the search process, accepting inferior solutions with probability to avoid local optima. At the same time, the adaptive multi-step neighborhood generation algorithm is used to intelligently generate neighborhood solutions to accelerate convergence and ensure that the optimal solution is obtained.

[0131] Example 4 To systematically evaluate the model's generalization ability in real-world application scenarios, this embodiment verifies and analyzes the above-mentioned site selection method for reservoir resettlement areas.

[0132] Reference Figure 6 As shown, this embodiment identifies 310 existing residential settlements within the land acquisition area. Within a given administrative region, and considering land use planning data and local natural resources, economic and social data, 10 candidate resettlement areas meeting the constraints are selected. Figure 7 ).

[0133] The first step is to determine the constraints on the total area of ​​the resettlement area and to determine whether the administrative region can meet the conditions for reservoir migrant resettlement.

[0134] The second step is to construct the solution space matrix, and then... The elements form a 0-1 matrix (310 rows, 10 columns) representing the site selection and planning scheme for the reservoir resettlement area. The representation is as follows: ; Next, we define the solution space as all matrices that satisfy the following conditions: ; The third step is to determine the number of resettlement areas, and to make a preliminary plan based on spatial data and socio-economic information. The objective function here is to minimize the number of proposed resettlement areas.

[0135] By verifying flood control and security constraints, spatial capacity constraints, and individual cost constraints, a feasible solution has been determined. If a feasible solution cannot be obtained, for ; until no feasible solution can be found.

[0136] The fourth step involves a hybrid approach of adaptive neighborhood generation and intelligent algorithms to find the optimal solution. The feasible solution from the third step is used as the initial solution, and the simulated annealing algorithm, which is the example method in this embodiment, is employed to perform a global search. See Embodiment 1 above for details, which will not be repeated here.

[0137] The fifth step is to perform a feasibility test on each generated neighborhood solution. On one hand, it must meet the topological matrix constraints of this invention; on the other hand, it must also meet the constraints and verification conditions of the location selection model. If the feasibility test passes, the neighborhood solution is taken as the new solution. If the feasibility test fails, it is regenerated.

[0138] The sixth step is to decide whether to accept the new solution based on the current solution, and to update the probability model and algorithm parameters.

[0139] Define the energy function The decision on whether to accept the new solution is based on the energy difference between the new and old solutions.

[0140] The methods and systems used in this embodiment are for updating the probability model and algorithm parameters, including but not limited to: row and column selection, simulated annealing parameters, and operation step distribution.

[0141] In addition, in the current If the adaptive neighborhood generation-intelligent algorithm hybrid optimal solution process fails to find a feasible solution, it is fed back to the modified solution process for controlling the number of resettlement areas. := Process. Continue until the optimal solution can be obtained. This is the minimum value, and the corresponding solution X is the desired value.

[0142] Step 7: Once the termination condition is met, the optimal solution is obtained, and the solution is output.

[0143] The optimal solution obtained by the adaptive neighborhood generation-intelligent algorithm hybrid solution method of this invention is mapped to the site selection and relocation scheme of the reservoir resettlement area, such as... Figure 8 .

[0144] In this way, the determined administrative area, based on land use planning data and using intelligent algorithms, achieves the following objectives: minimizing the total distance for relocation and resettlement of all residents; meeting the risk target of ensuring that the flood control standards of the proposed resettlement area meet the prescribed defense standards; and ensuring fairness by considering construction costs and long-term management while preserving the original customs and features of the resettled residents without causing division.

[0145] This application also provides a computer-readable storage medium storing a computer program. When executed by a processor, the computer program implements the method in any of the embodiments of this application. Specifically, a system or apparatus equipped with a storage medium may be provided, on which software program code implementing the functions of any of the above embodiments is stored, and the computer (or CPU (Central Processing Unit) or MPU (Microprocessor Unit) of the system or apparatus may read and execute the program code stored in the storage medium.

[0146] Specifically, according to embodiments of this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this application include a computer program product comprising a computer program carried on a computer-readable storage medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication component, and / or installed from a removable medium. When the computer program is executed by a central processing unit (CPU), it performs the functions defined above in the system of this application.

[0147] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram or flowchart, and combinations of blocks in a block diagram or flowchart, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0148] The units described in the embodiments of this application can be implemented in software or hardware, and the described units can also be located in a processor. The names of these units do not necessarily limit the specific unit itself.

[0149] It should be noted that although several modules or units for the device used to perform actions have been mentioned in the detailed description above, this division is not mandatory. In fact, according to the embodiments of this application, the features and functions of two or more modules or units described above can be embodied in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided and embodied by multiple modules or units.

[0150] In the several embodiments provided in this application, it should be understood that the disclosed systems, modules, and methods can be implemented in other ways. For example, the module embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces, or indirect coupling or communication connection between modules or units, and may be electrical, mechanical, or other forms.

[0151] The above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit it. This application is not limited to the exact structures described above and illustrated in the accompanying drawings, and it should not be considered that the specific implementation of this application is limited to these descriptions. For those skilled in the art, various changes and modifications made without departing from the concept of this application should be considered to fall within the protection scope of this application.

Claims

1. A method for selecting a site for a reservoir resettlement area, characterized in that, include: Based on the spatial data and socio-economic information corresponding to the target study area, as well as the requirements for reservoir resettlement areas, a multi-objective optimization planning model for reservoir resettlement areas is constructed, and an initial site selection planning scheme is generated. The initial site selection planning scheme is represented topologically to obtain the existing residential gathering point-resettlement area allocation matrix based on 0-1 matrix form, and the solution space is constructed; wherein, each element in the existing residential gathering point-resettlement area allocation matrix represents the allocation relationship between residential points and resettlement areas, presenting the mapping from existing residential gathering points to resettlement areas; The initial number of proposed resettlement areas was estimated. And randomly designate within the target study area One alternative resettlement area; Based on the solution space, a simulated annealing algorithm is used to perform a global search, and inferior solutions are accepted with probability to avoid local optima. This is combined with the existing residential cluster-resettlement area allocation matrix and the requirement for a certain number of resettlement areas. The constrained adaptive multi-step neighborhood generation algorithm iteratively optimizes and adaptively adjusts the search strategy under feasibility constraints, and converges to obtain the optimal resettlement area site selection scheme. Specifically, based on the solution space, a simulated annealing algorithm is used to perform a global search, and inferior solutions are accepted with probability to avoid local optima. Simultaneously, combined with the existing resident gathering point-resettlement area allocation matrix and an adaptive multi-step neighborhood generation algorithm that satisfies the constraint on the number of resettlement areas, the search strategy is iteratively optimized and adaptively adjusted under feasibility constraints to converge the optimal resettlement area location scheme, including: Define the energy function Calculate the energy change caused by a single point movement. Wherein, the energy function Used to determine the quality of the current solution; The adaptive multi-step neighborhood generation algorithm and the simulated annealing algorithm are initialized with parameters, including probability model initialization, operation step distribution initialization, simulated annealing parameter initialization and solution initialization; Select the row to be adjusted based on the initialized probability model. and target column Perform a single-point movement operation to generate a neighborhood solution. And in the solution of the neighborhood After feasibility verification, multi-step operations are performed according to the initial step distribution to form a composite neighborhood, and the probabilistic model is updated based on the quality of the generated neighborhood solutions. Then, based on the updated probabilistic model obtained after the initial neighborhood solution calculation, non-initial neighborhood solutions are calculated. The content of the non-initial neighborhood solutions includes: For the current solution X, determine the row corresponding to the maximum row selection probability based on the probability model updated after the initial calculation of the neighborhood solution. And the column corresponding to the uniform distribution of the maximum target column selection probability. After performing a single-point movement operation, a new neighborhood solution is generated. After the feasibility of the new neighborhood solution is verified, the number of operations to generate a new neighborhood solution is taken as the number of steps corresponding to the step number distribution probability with the highest probability. Multi-step operations are performed to form a composite neighborhood and the probability model is updated according to the quality of the generated new neighborhood solution. Furthermore, during the algorithm iteration process, the non-first neighborhood solution is repeatedly executed, and a new solution is generated each time based on the probability model updated in the previous round, until the termination condition is met and the optimal solution that meets the constraints is obtained. For each newly generated solution, a simulated annealing acceptance criterion is used to determine whether the new solution can be accepted. If the new solution is accepted, the probability model is updated based on the quality of the new solution until the termination condition is met.

2. The method for selecting a site for a reservoir resettlement area according to claim 1, characterized in that, The reservoir resettlement area planning model includes a site selection objective function and site selection constraints; the site selection objective function includes a first objective function and a second objective function, the first objective function is to minimize the sum of weighted distances from all existing residential settlements to the proposed resettlement area location, and the second objective function is to minimize the number of proposed resettlement areas; The address selection objective function is specifically as follows: ； ； ； In the formula, This indicates the number of alternative resettlement areas. This indicates the current number of residential settlements. This represents the population size coefficient of immigrants relocated from existing residential areas. The weighting of resettlement preference is determined by the degree of centralized resettlement. Let be the decision variable, representing the current residential gathering points. Assigned to the designated resettlement area The plan; among which, when The time indicates that the first The current residential settlements' immigrants were allocated to the first... Resettlement will be carried out in the reservoir resettlement area. At that time, it was indicated that the first The current residential settlements' immigrants were allocated to the first... A reservoir resettlement area for displaced people; For state description, Indicates except All other allocations besides those; Indicates the current residential gathering points to the proposed resettlement area Distance coefficient, Indicates the current residential gathering points to the proposed resettlement area The traffic cost coefficient; The site selection constraints include comprehensive coverage constraints, flood control and security constraints, spatial capacity constraints, and distance limitation constraints; wherein, the comprehensive coverage constraint specifically refers to: ； ； ； In the formula, This indicates the number of alternative resettlement areas. This indicates the current number of residential settlements; Indicates the current residential gathering points Assigned to the designated resettlement area The plan; Indicates the number of proposed resettlement areas; Indicates the determination of alternative resettlement areas Whether to include it as a variable in the proposed resettlement area, when considering alternative resettlement areas When as a proposed resettlement area =1, otherwise =0; The flood control and security constraints are specifically as follows: ； ； ； in: Resettlement area The elevation corresponding to the flood control standard is given by , and n represents the return period corresponding to the flood control standard. This represents the elevation corresponding to the flood defense standard with a return period of n. This indicates that the planned relocation and resettlement will be carried out in the alternative resettlement area. The number of immigrants, This represents the population size coefficient of immigrants relocated from existing residential areas. The weighting of resettlement preference is determined by the degree of centralized resettlement. Indicates the current residential gathering points Assigned to the designated resettlement area The plan This indicates the current number of residential settlements; The spatial capacity constraint refers to the number of immigrants allocated to a candidate resettlement area not exceeding the capacity of the candidate resettlement area and also satisfying the area constraint of the candidate resettlement area. The distance restriction constraint is specifically as follows: ； ； ； in, This refers to mandatory distance restrictions determined based on the actual circumstances of the administrative region. This indicates consideration of the current residential gathering points. The weighted distance limit value is derived after considering topography, economic factors, and social factors. Indicates the current residential gathering points The relocation economic parameters generated by the topography and landforms Indicates the current residential gathering points i The economic parameters of relocation generated by the economy, Indicates the current residential gathering points The relocation economic parameters generated by social factors, Indicates the current residential gathering points to the proposed resettlement area Distance coefficient, This indicates the number of alternative resettlement areas. This indicates the current number of residential settlements. Indicates the determination of alternative resettlement areas Whether to be considered as a variable for the resettlement area Indicates the current residential gathering points Assigned to the designated resettlement area The plan It represents an economic coefficient that measures the costs, primarily related to relocation, resulting from the influence of distance, natural geography, and socio-economic conditions.

3. The method for selecting a site for a reservoir resettlement area according to claim 2, characterized in that, The initial site selection planning scheme is topologically represented to obtain a current residential gathering point-resettlement area allocation matrix based on a 0-1 matrix, and the solution space is constructed, including: Each As an element, it manifests in the reservoir resettlement area. OK 0-1 matrix of columns ,in, There is one and only one 1 in the middle row; Based on the 0-1 matrix Define the solution space The solution space Specifically: ； The 0-1 matrix is ​​determined by considering the fixed number of existing residential gathering points and the fixed number of planned resettlement areas in the current stage. At least one row of 1s is randomly moved, and the number of columns whose cumulative sum is not zero remains constant. ; For the solution space For each solution X, the corresponding column and vector are... Each component And based on columns and vectors, calculate the set of zero columns. and the number of zero columns This transforms the activation status of the resettlement area into quantifiable matrix features.

4. The method for selecting a site for a reservoir resettlement area according to claim 3, characterized in that, In the reservoir resettlement area planning model, the process of generating neighborhood solutions based on the current solution must satisfy the following conditions: There is at least one row , making =1, becomes =1, and j≠k; Number of zero columns in neighborhood solutions With the current solution Number of zero columns Maintain consistency; Solutions in any two solution spaces can be converted to each other through a finite number of neighborhood operations; In the process of generating neighborhood solutions, for the current fixed-dimensional 0-1 matrix form, the basic operation is defined as: selecting rows. and target column , will be done 1 from column Move to column Specifically: ； ； ； ； To set feasibility conditions for the current column and s, one of the following conditions, A and B, must be met: ； In the formula, To perform a single-point movement operation on the 0-1 matrix of the first element Line 1 The new value of the column, Represents the 0-1 matrix. Line 1 The position of the column corresponds to the current residential gathering point. The originally allocated proposed resettlement area ; Represents the 0-1 matrix. Line 1 The position of the column corresponds to the current residential gathering point. Proposed resettlement areas for relocation ; Indicates the division of a 0-1 matrix by... and The remaining positions outside; For the first 0-1 matrix before the move operation Line 1 The original values ​​of the column; For the original resettlement area after the relocation operation The corresponding column and the new value, Original resettlement area before relocation operation The corresponding columns and; For the original resettlement area after the relocation operation The corresponding column and the new value, Original resettlement area before relocation operation The corresponding columns and; After the move operation Resettlement areas in any other location The corresponding column and the new value, Pre-operation placement area The columns and; This indicates a logical AND.

5. The method for selecting a site for a reservoir resettlement area according to claim 4, characterized in that, The process of using simulated annealing acceptance criteria to determine whether a new solution can be accepted for each generated solution includes: Determine whether the energy value of the new solution meets the preset maximum energy value constraint; If the energy value of the new solution meets the constraint condition of the maximum energy value, calculate the energy change value corresponding to the new solution, and apply the simulated annealing acceptance criterion to decide whether to accept the new solution based on the energy change value; wherein, the simulated annealing acceptance criterion is: ； In the formula, It is the current temperature and energy change. ;if If the solution is correct, then accept the new solution; otherwise, use probability. Accept the new interpretation.

6. The method for selecting a site for a reservoir resettlement area according to claim 5, characterized in that, The quality update probability model based on the new solution includes: The probabilistic model is updated using feedback information from historical searches; specifically, this includes: adaptive updates of row selection probability, column selection probability, step distribution, inverse temperature parameter, and simulated annealing cooling temperature, among which: The adaptive update of the row selection probability is as follows: ； ； ； In the formula, Indicates line The probability of choosing, Indicates line The quality estimate; The learning rate represents the quality of the row and controls the degree of influence of new information; The inverse temperature parameter selected for row selection is used to control exploration and utilization. The larger the value, the more likely the row with the higher quality estimate will be selected. The smaller the size, the more uniform the distribution; This indicates the reward generated from this move; This indicates assignment; The adaptive update of column selection probability is as follows: ； ； in, Indicates the selected row In this case, select column The conditional probability; Indicates the inverse temperature parameter selected in the column; Indicates that will be carried out Move 1 to column The quality estimate; Indicates line i The set of feasible target columns; Indicates the traversal index in the rowizable set; The learning rate indicates the quality of the column; This indicates the reward generated from this move; This indicates assignment; The adaptive update of the step distribution is as follows: ； ； ； in, Indicates the number of steps selected. The probability of performing a neighborhood operation; The adjustment factor selected for the number of steps; Indicates the number of steps to traverse the index; Indicates the maximum allowed number of steps for neighborhood operations; Indicates the number of steps The quality estimate; The learning rate represents the quality of the steps. This represents the total reward for the entire sequence after a single multi-step operation. Indicates the first Energy difference in step operation; Indicates the first The fuzzy annealing temperature during step-by-step operation; The adaptive update of the inverse temperature parameter is as follows: ； In the formula, , They represent the first Second-rate, The inverse temperature parameter generated after the next neighborhood solution is generated. This represents the variance of all row quality estimates. This represents the mean of the quality estimates. ε It is a very small positive number that ensures the denominator is not zero. Indicates the first The fuzzy annealing temperature during the step operation Indicates the initial temperature; The update of the simulated annealing cooling temperature is as follows: ； In the formula, Indicates the first The temperature of the next time, T k Indicates the first k Temperature at the next iteration Represents the energy function. Represents the best energy in history. This refers to the parameter that controls the decay rate.

7. The method for selecting a site for a reservoir resettlement area according to claim 6, characterized in that, The termination condition is that the current temperature of the iteration is less than or equal to the preset minimum temperature, or the improvement stagnation condition is reached, or the entropy of the probability model is less than the preset threshold.

8. The method for selecting a site for a reservoir resettlement area according to any one of claims 6-7, characterized in that, Also includes: Step 5.1: If, after solving for the optimal solution based on the initial proposed number of resettlement areas, a feasible solution that meets the constraints cannot be generated, the following adjustments are made: Among the alternative resettlement areas The allocation of proposed resettlement areas is then recalculated to find the optimal resettlement area location scheme. Step 5.2, in adjustment Among the alternative resettlement areas The allocation of proposed resettlement areas is recalculated, and the optimal resettlement area site selection scheme is solved again. If a feasible solution is obtained, adjustments are made step by step to make the optimal resettlement area site selection scheme more feasible. Then, in the adjusted allocation scheme, the candidate resettlement areas with the assigned values ​​are randomly selected as the new allocation scheme for the proposed resettlement areas. The optimal resettlement area location scheme is solved until no feasible solution can be obtained. Then, the current number of proposed resettlement areas is determined as the final number of proposed resettlement areas. Step 5.3, in adjustment Among the alternative resettlement areas If the allocation of the proposed resettlement areas is repeated, and the optimal resettlement area site selection scheme cannot be obtained, then let... And randomly assign values ​​within the target study area. After selecting one alternative resettlement area, the optimal resettlement area location scheme is solved again. If the optimal solution cannot be generated based on the number of proposed resettlement areas after assignment, and if no feasible solution that meets the constraints can be generated, steps 5.1 to 5.3 are executed until the optimal resettlement area location scheme is obtained.

9. A site selection system for reservoir resettlement areas, used to implement the site selection method for reservoir resettlement areas according to any one of claims 1-8, characterized in that, include: The initial scheme generation module is used to construct a multi-objective optimization planning model for reservoir resettlement areas based on the spatial data and socio-economic data corresponding to the target research area, as well as the requirements for reservoir resettlement areas, and to generate initial site selection planning schemes. The solution space construction module is used to perform topological representation of the initial site selection planning scheme, obtain the current residential gathering point-resettlement area allocation matrix based on 0-1 matrix form, and construct the solution space; wherein, each element in the current residential gathering point-resettlement area allocation matrix represents the allocation relationship between residential points and resettlement areas, presenting the mapping from current residential gathering points to resettlement areas; The optimal solution module is used to perform a global search based on the solution space using the simulated annealing algorithm, and to accept inferior solutions with probability to avoid local optima. At the same time, it combines the existing residential gathering point-relocation area allocation matrix and the adaptive multi-step neighborhood generation algorithm that satisfies the constraint of the number of resettlement areas, to iteratively optimize under the feasibility constraint and adaptively adjust the search strategy to converge to obtain the optimal resettlement area site selection scheme. Specifically, based on the solution space, a simulated annealing algorithm is used to perform a global search, and inferior solutions are accepted with probability to avoid local optima. Simultaneously, combined with the existing resident gathering point-resettlement area allocation matrix and an adaptive multi-step neighborhood generation algorithm that satisfies the constraint on the number of resettlement areas, the search strategy is iteratively optimized and adaptively adjusted under feasibility constraints to converge the optimal resettlement area location scheme, including: Define the energy function Calculate the energy change caused by a single point movement. Wherein, the energy function Used to determine the quality of the current solution; The adaptive multi-step neighborhood generation algorithm and the simulated annealing algorithm are initialized with parameters, including probability model initialization, operation step distribution initialization, simulated annealing parameter initialization and solution initialization; Select the row to be adjusted based on the initialized probability model. and target column Perform a single-point movement operation to generate a neighborhood solution. And in the solution of the neighborhood After feasibility verification, multi-step operations are performed according to the initial step distribution to form a composite neighborhood, and the probabilistic model is updated based on the quality of the generated neighborhood solutions. Then, based on the updated probabilistic model obtained after the initial neighborhood solution calculation, non-initial neighborhood solutions are calculated. The content of the non-initial neighborhood solutions includes: For the current solution X, determine the row corresponding to the maximum row selection probability based on the probability model updated after the initial calculation of the neighborhood solution. And the column corresponding to the uniform distribution of the maximum target column selection probability. After performing a single-point movement operation, a new neighborhood solution is generated. After the feasibility of the new neighborhood solution is verified, the number of operations to generate a new neighborhood solution is taken as the number of steps corresponding to the step number distribution probability with the highest probability. Multi-step operations are performed to form a composite neighborhood and the probability model is updated according to the quality of the generated new neighborhood solution. Furthermore, during the algorithm iteration process, the non-first neighborhood solution is repeatedly executed, and a new solution is generated each time based on the probability model updated in the previous round, until the termination condition is met and the optimal solution that meets the constraints is obtained. For each newly generated solution, a simulated annealing acceptance criterion is used to determine whether the new solution can be accepted. If the new solution is accepted, the probability model is updated based on the quality of the new solution until the termination condition is met.

Images