Method, device and storage medium for generating long-term production plan scheme of open-pit multi-mining area combined mining
By using a block-level mixed integer programming model and a two-stage optimization framework, a long-term production plan for joint open-pit mining of multiple mining areas was generated, which solved the problem of mutual influence between mining areas and achieved efficient production optimization and improved economic benefits.
Patent Information
- Application Number
- CN202611150372.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-07-31
- Publication Date
- 2026-11-03
AI Technical Summary
Existing technologies fail to fully reflect the mutual influence between mining areas in the long-term planning optimization of joint open-pit mining of multiple mining areas, resulting in shared capacity conflicts, excessive concentrate inventory, insufficient order delivery, fluctuations in the grade of processed ore, and delays in the mining of high-value resources. Furthermore, large-scale combination optimization problems are difficult to solve within a reasonable timeframe.
A mixed-integer programming model at the block level is adopted, which combines the prior relationships and spatial adjacency of blocks. A candidate mining plan list is generated through the mining body sequence and planning period configuration. A two-stage collaborative optimization framework is used to perform global optimization and local fine-tuning, so as to achieve unified optimization of production capacity across mining areas.
It effectively solves the problem of high solution complexity of large-scale block-level optimization models, generates high-quality long-term production plans for joint open-pit mining of multiple mining areas, and improves economic efficiency and production balance.
Smart Images
Figure CN123032045A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to intelligent mining planning technology, and more particularly to a method for generating a long-term production plan for joint open-pit mining of multiple mining areas. Background Technology
[0002] Optimizing long-term open-pit mine mining plans involves combining ore deposit block models, 3D spatial information, and computer optimization technology. Under constraints such as slope safety, mine production capacity, and ore processing capacity, it optimizes the sequence of ore extraction, determines the mining and stripping locations, scales, and supply schemes for each planning period, and is a crucial prerequisite for achieving intelligent production and efficient resource utilization in open-pit mines. Secondly, as mine production scales expand, multiple open-pit mines sharing ore supply tasks is becoming an important production organization model. Unified optimization of the mining sequence, production capacity, concentrate inventory, order delivery, and blending requirements across multiple mines can fully utilize the differences in resource grade, ore value, and exploitable space among different mines, ensuring balanced and stable mine production while improving the economic benefits throughout the project's entire lifecycle.
[0003] Currently, in long-term planning optimization studies for joint open-pit mining of multiple mining areas, mining plans for each mining area are typically prepared separately, and then manually coordinated based on the company's overall mining and stripping capacity and ore processing capacity. This method fails to fully reflect the mutual influence of different mining areas in terms of shared production capacity, concentrate inventory, order delivery, and ore blending requirements. It also struggles to unify the pre-existing relationships of ore blocks within a mining area with the resource allocation relationships between mining areas, easily leading to shared capacity conflicts, excessive concentrate inventory, insufficient order delivery, fluctuations in the grade of processed ore, and delays in the mining of high-value resources. Secondly, the long-term planning optimization problem for joint open-pit mining of multiple mining areas is a large-scale combinatorial optimization problem. As the number of mining areas, ore blocks, and planning periods increases, the number of decision variables and constraints in the ore block-level mixed integer programming model grows rapidly, making it difficult to solve effectively within a reasonable timeframe using general heuristic algorithms or mathematical programming solvers.
[0004] Therefore, a high-quality long-term plan for joint open-pit mining in multiple mining areas is needed, which can take into account both the spatial advancement of mining areas and the sharing of production capacity. Summary of the Invention
[0005] The technical problem to be solved by the present invention is to provide a method, equipment and storage medium for generating a long-term production plan for joint open-pit mining of multiple mining areas, in order to address the deficiencies in the prior art.
[0006] The technical solution adopted by this invention to solve its technical problem is: a method for generating a long-term production plan scheme for joint open-pit mining in multiple mining areas, comprising the following steps: 1) Collect production and planning parameters for multiple open-pit mines, including: block models, block attributes, spatial locations, and prior relationships of multiple open-pit mines; 2) Based on the block model, block attributes, and block prerequisites, and combined with economic parameters, establish a block-level mixed integer programming reference model; The reference model for mixed-integer programming at the block level aims to maximize the net present value of a project. The constraints of the model include block preconditions, unique mining constraints, production capacity constraints, inventory constraints, order delivery constraints, and ore blending constraints. 3) For each open-pit mine area, the ore block is divided into multiple mining bodies according to the prior relationship of the ore blocks, the spatial adjacency relationship and the preset minimum mining increment; 4) Generate a sequence of candidate mining bodies based on the pre-existing relationship between the mining bodies, and configure a planning period for each mining body in the candidate mining body sequence to form multiple candidate mining plan columns, wherein each candidate mining plan column contains the mining body sequence and planning period configuration of the mining blocks required to be mined in the corresponding open-pit mining area; 5) Extract the attributes of each candidate mining plan column, including the phased mining and stripping volume, the amount of ore fed into the beneficiation process, the concentrate output, the grade, and the discounted cost, and construct a multi-choice set to divide the constrained master problem in order to select the best single candidate column for each mining area, and uniformly constrain cross-mining area sharing capabilities, concentrate inventory, order delivery, and blending grade. 6) Establish a two-stage collaborative optimization framework to achieve model solution that combines global optimization with local fine-tuning; The first stage is global heuristic optimization; the second stage is local refined reconstruction. 7) Split the reconstructed joint production scheduling solution by mining area and recombin it with the unmodified part of the original candidate column to form a feedback candidate mining plan column; add the feedback column that meets the requirements of local feasibility, no duplication and has positive reduction benefits or comes from a better integer solution to the candidate column set, and return to step 6) for iterative optimization; 8) After the iterative optimization reaches the termination condition, solve the final integer-constrained master problem and output the long-term production plan scheme for joint mining of multiple open-pit mines.
[0007] According to the above scheme, in step 2), the objective function of the block-level mixed integer programming reference model is:
[0008] In the formula, For the planning period set; It is a collection of concentrates or resources; For the planning period The discount factor; For the planning period Internal resources The unit delivery value of the corresponding concentrate; For the planning period Internal resources The actual delivery volume of the corresponding concentrate; For the planning period The costs of mining, transporting, and processing all the mineral blocks within the mine; For initial investment.
[0009] According to the above scheme, in step 2), the model construction process includes: determining the concentrate output corresponding to the block based on the block quality, ore or waste rock attributes, resource grade, beneficiation recovery rate and concentrate grade; determining the block operation cost based on the block quality, mining cost, beneficiation cost, ore transportation cost and waste rock transportation and disposal cost; determining the concentrate delivery revenue based on the actual concentrate delivery volume and unit delivery value for each planning period; and establishing the block-level mixed integer programming model with the decision variable of whether the block is mined within the corresponding planning period and the goal of maximizing the project's net present value.
[0010] According to the above scheme, the pre-mining constraint of the ore block is used to ensure that the pre-mining ore block corresponding to the ore block scheduled for mining is mined in the same planning period or an earlier planning period. The unique mining constraint of the block is used to ensure that each block is mined at most once during the entire planning period, and to ensure that the blocks required to be mined within the final boundary are mined during the entire planning period. The production capacity constraints include single-mine area mining and stripping capacity constraints, company-wide overall mining and stripping capacity constraints, and mineral processing capacity constraints. The single mining area stripping capacity constraint is used to limit the stripping volume of any open-pit mine in any planning period to not exceed the stripping capacity of that open-pit mine. The company's overall mining and stripping capacity constraint is used to limit the total mining and stripping volume of all open-pit mines within the same planning period from not exceeding the company's overall mining and stripping capacity. The mineral processing capacity constraint is used to limit the total amount of ore fed into all open-pit mines within the same planning period to no more than the mineral processing capacity.
[0011] According to the above scheme, in step 3), the method for dividing the mining body includes: The set of currently mineable blocks is determined based on the blocks that have not yet been included in the mining body and their preceding relationships. Select a seed block from the current set of mineable blocks. Blocks that are spatially adjacent to the current set of blocks and whose predecessor blocks are already located in the previous mining body or the current set of blocks are designated as expandable blocks. The expansion order is determined according to economic value, spatial distance, direction of advance, or resource grade until the minimum mining increment is reached.
[0012] According to the above scheme, in step 5), the column coefficients are determined based on the phased attributes of the candidate mining plan column, and the candidate column selection variables are used to construct a multi-selection set to divide the restricted principal problem.
[0013] The construction of the restricted master problem for multiple choice set partitioning includes the following steps: (1) Establish a set of candidate mining plans for each open-pit mine area, and calculate the mining and stripping volume, ore input volume, resource grade weighted volume and concentrate output for each planning period according to the planning period of the ore blocks in each candidate mining plan, as well as the discounted operating cost of the candidate mining plan. (2) Set a candidate column selection variable to indicate whether the candidate mining plan column of the corresponding open-pit mine area is selected; (3) Set variables for actual concentrate delivery quantity and concentrate ending inventory quantity; (4) The objective is to maximize the net present value of the project after deducting the discounted operating costs of the selected candidate mining plans from the actual revenue delivered by the concentrate in each planning period; (5) Establish candidate column selection constraints that ensure that each open-pit mine area selects exactly one candidate mining plan column; (6) Based on the phased attributes of the selected candidate mining plan, establish constraints on single mining area stripping capacity, overall company stripping capacity, mineral processing capacity, concentrate inventory, order delivery and blending grade.
[0014] According to the above scheme, in step 6), Phase 1: Generate initial joint production scheduling solutions and initial candidate column sets that satisfy full constraints for single mining area and cross mining area; solve the linear relaxation master problem to obtain dual information, use the reduction benefit as a screening index, continuously iterate to generate high-quality new candidate columns, and after the iteration reaches the standard, solve the integer-constrained master problem to obtain the basic optimal solution; The second stage involves calculating the capacity utilization rate and the proximity of inventory to grade boundaries for each planning period to screen key planning periods. An adaptive large neighborhood search algorithm is then used to reconstruct the mining arrangement for key periods by using destruction and repair operators. The weights are dynamically updated based on the historical optimization effects of the operators to achieve refined optimization of local solutions.
[0015] The present invention also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described scheme.
[0016] The present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the method in the above-described scheme.
[0017] The beneficial effects of this invention are: 1. Based on the pre-existing relationship of ore blocks, spatial adjacency, and minimum mining increment, this invention aggregates scattered ore blocks into standardized mining bodies. It replaces massive ore block-level decision variables with mining body sequences and candidate columns configured during the planning period, which greatly reduces the scale of the main problem solution and effectively solves the problem of high complexity and long time consumption in solving large-scale ore block-level optimization models in open-pit multi-mine areas. 2. By splitting, reorganizing and reconstructing the production scheduling solution, a brand-new feedback column is generated. The verified high-quality feedback column is continuously included in the candidate set to expand the candidate solution space and improve the planning economy. Attached Figure Description
[0018] The present invention will be further described below with reference to the accompanying drawings and embodiments. In the accompanying drawings: Figure 1 This is a flowchart of a method according to an embodiment of the present invention; Figure 2 This is a schematic diagram of the principle of reconstructing the multi-choice set partitioning restricted master problem using a block-level mixed integer programming reference model according to an embodiment of the present invention; Figure 3 This is a flowchart of the two-stage collaborative optimization and feedback candidate column update process of heuristic column generation and adaptive large neighborhood search in an embodiment of the present invention. Detailed Implementation
[0019] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.
[0020] like Figure 1 As shown, a method for generating a long-term production plan for joint open-pit mining of multiple mining areas includes the following steps: S1) Obtain block models, spatial locations, block relationships, economic parameters, production capacity, concentrate inventory, order delivery, ore blending and planning parameters for multiple mining areas; In this embodiment, it is assumed that all ore blocks have the same mass, and the mass of each ore block is set to be... The spatial dimensions, quality, and properties of the ore blocks remain unchanged during the planning period, and each ore block is mined at most once during the entire planning period. The prior relationships between ore blocks are determined based on the open-pit mine slope angle, bench height, and spatial location of the ore blocks, forming a directed acyclic graph.
[0021] S2) Construct a block-level mixed-integer programming reference model; With maximizing the project's net present value as the core objective, a benchmark optimization model is built based on the collected fundamental parameters; the establishment of the block-level mixed integer programming reference model includes: The concentrate output coefficient of each ore block is determined based on the ore or waste rock properties, resource grade, beneficiation recovery rate and concentrate grade. The operating cost of each ore block is determined based on the ore block quality, mining cost, ore transportation cost, waste rock transportation and disposal cost, and mineral processing cost. Whether the ore block is mined during the corresponding planning period is used as the binary ore block mining decision variable, and the actual delivery volume and ending inventory of various concentrates during the corresponding planning period are used as continuous decision variables. Based on the mining decisions of all open-pit mines in each planning period, determine the single-mine stripping volume, the company's overall stripping volume, the amount of ore fed into the beneficiation process, the concentrate output, and the grade-weighted amount of the beneficiated ore.
[0022] At the same time, constraints are set, including constraints on pre-mining of ore blocks, constraints on unique mining of ore blocks, constraints on production capacity, constraints on concentrate inventory, constraints on order delivery, and constraints on blending grade, to form a standardized ore block-level optimization benchmark model.
[0023] Let the open-pit mine area be set as mining area The collection of mineral blocks is The resource collection is The planning period set is .set up For ore discrimination parameters, when the mining area mineral blocks When it is ore, When it is waste rock, ; For mineral blocks China Resources The grade of the raw ore; For resources ore beneficiation recovery rate; For resources The grade of the concentrate.
[0024] Mineral Block China Resources The corresponding concentrate yield coefficient per unit mass is:
[0025] Therefore, the mineral block Resources generated after being mined and selected The corresponding concentrate quantity is .
[0026] set up The cost of mining per unit mass of ore block. For ore transportation costs, Costs associated with waste rock transportation and disposal. If the cost of beneficiation per unit mass of ore is given, then the operating cost of the ore block is:
[0027] set up Indicates mining area mineral blocks Is it in the planning period? Mining. When the ore block is mined during the planning period. ,otherwise .set up For the planning period resource The actual delivery volume of the corresponding concentrate. The value delivered per unit of the corresponding concentrate. The discount rate is... For the initial investment of the project, the objective function of the block-level mixed integer programming model is: ; The objective function described above only includes the actual delivered concentrate in revenue, while inventory concentrate is not included in the current period's revenue.
[0028] The constraints are as follows: Pre-mining constraints are used to ensure that the preceding blocks corresponding to the blocks scheduled for mining are mined in the same planning period or an earlier planning period. The unique mining constraint for each block is designed to ensure that each block is mined at most once during the entire planning period, and that blocks requiring mining completion within the final boundary are mined within the entire planning period.
[0029] Production capacity constraints include single-mine stripping capacity constraints, overall company stripping capacity constraints, and mineral processing capacity constraints. The single-mine stripping capacity constraint is used to limit the stripping volume of any open-pit mine in any planning period to not exceed the stripping capacity of that open-pit mine; the overall company stripping capacity constraint is used to limit the sum of the stripping volumes of all open-pit mines in the same planning period to not exceed the overall company stripping capacity; the mineral processing capacity constraint is used to limit the sum of the amount of ore fed into the mineral processing area of all open-pit mines in the same planning period to not exceed the mineral processing capacity.
[0030] The inventory constraint determines the final concentrate inventory for the current planning period based on the concentrate production and delivery volume of the current planning period, and limits the final concentrate inventory to a preset range. For the first planning period, the inventory constraint makes the final inventory equal to the sum of the initial concentrate inventory and the concentrate production for the current planning period minus the actual delivery volume for the current planning period. For subsequent planning periods, the concentrate inventory constraint makes the final inventory equal to the sum of the final inventory of the previous planning period and the concentrate production for the current planning period minus the actual delivery volume for the current planning period, and keeps the final inventory between the corresponding lower and upper inventory limits for concentrate. Order delivery constraints ensure that the actual delivery volume of various concentrates within each planning period falls between the lower and upper limits of the corresponding order delivery volume. The ore blending constraint ensures that the grade weighted amount of various resources and the amount of ore fed into the beneficiation process satisfy a linear grade relationship, so that the average grade of the ore fed into the beneficiation process is between the corresponding lower and upper grade limits.
[0031] set up For mineral blocks The set of preceding mining blocks. Mining blocks Before it is mined, its preceding ore blocks should be mined within the same planning period or an earlier planning period, therefore: ; Each block may be mined at most once during the entire planning period: ; For the set of mineral blocks within the final realm that are required to be mined within the planning period. ,satisfy: ; set up For the mining area During the planning period Mining capacity, For the company's overall mining and stripping capabilities, For the mineral processing capacity, we have: ; ; ; in, For the planning period The total amount of ore entering the concentrator in all mining areas.
[0032] resource During the planning period Taste weighted quantity and concentrate production They are respectively: ; ; Concentrate inventory is balanced based on previous period inventory, current period production and current period delivery; the initial inventory for the first period is the aforementioned initial inventory level.
[0033] ; ; in, For concentrate inventory, This is the lower limit of inventory. This is the upper limit for inventory.
[0034] set up and Let the minimum and maximum delivery quantities be the quantities specified in the order, respectively. Then the order requirement constraints are: ; When an order requires a fixed delivery quantity, .
[0035] set up and respectively planning period Resources in selected ores Given the lower and upper limits of the ore grade, the blending requirements are as follows: ; When the amount of ore fed into the process is greater than zero, the constraint limits the average grade; when the amount fed into the process is zero, the linear constraint still holds.
[0036] S3) Mining body divided into aggregated ore blocks; For each open-pit mine, after establishing a reference model, the ore blocks are divided into spatially continuous mining volumes based on the block prior relationships, spatial adjacency relationships, and minimum mining increments.
[0037] The minimum mining increment is set based on the differences in ore block quality: when the ore block quality is consistent, the total quality corresponding to the preset number of ore blocks is used as the benchmark; when the ore block quality is uneven, the cumulative quality is used as the benchmark. The generation of mining bodies follows feasibility rules: screening the set of mineable blocks, selecting seed blocks, expanding blocks according to dimensions such as economic value, grade, and spatial distance, and when the remaining small blocks cannot meet the standards, adjustments are made through adjacent merging or boundary backtracking to ensure that all mining bodies meet the minimum scale, spatial continuity, and pre-compliance requirements.
[0038] set up This is the minimum number of ore blocks corresponding to the preset minimum mining increment. Let the mass be a positive integer; let the mass of each ore block be . The minimum mining mass of the basic mining body for: ; The set of currently mineable blocks is determined based on the ore blocks that have not yet been included in the mining body and their preceding relationships; For mining area m, select seed blocks from it to establish the current block set P.
[0039] ; in: Generate an iteration number for the mining body; Indicates the first The set of ore blocks that have been included in the mining body after the next iteration; This represents the set of currently mineable blocks determined based on allocated blocks and prior relationships. For the current set of blocks that is expanding; The set of ore blocks contained in the preceding mining body that has already been formed.
[0040] Blocks that are spatially adjacent to the current block set P and whose preceding blocks are already located in the previous mining body set U or the current block set P are considered expandable blocks, and their expansion order is determined according to economic value, spatial distance, direction of advancement, or resource grade.
[0041] The expansion priority is determined based on at least one of the following: economic value of the ore block, spatial distance, mining area advancement direction, and resource grade. The expandable ore block is then added to the current ore block set. When the cumulative number or quality of the current block set reaches the preset minimum mining increment, the current block set is determined as a mining body; the blocks in the mining body are marked as allocated, the current mineable block set is updated, and the above steps are repeated.
[0042] When the cumulative number of blocks or the cumulative mass of any remaining unallocated block in the spatial connectivity component is less than the preset minimum mining increment, the spatial connectivity component is incorporated into one or more mining bodies that are spatially adjacent to it and satisfy the block prerequisite relationship. If no mining body meets the merging conditions, backtrack and adjust the ore block boundaries of one or more mining bodies adjacent to the spatial connectivity component to re-form mining bodies that respectively meet the preset minimum mining increment, spatial continuity and ore block precedence relationship; The prerequisite relationships between mining bodies are determined based on the prerequisite relationships of the ore blocks contained in each mining body, and the mining bodies are topologically sorted accordingly.
[0043] S4) Generate a list of candidate mining plans Based on the pre-constraint relationships of mining bodies in each mining area, a compliant mining body time sequence is generated; different planning period configurations are matched for each sequence to form a candidate mining plan list that includes the mining sequence of mining bodies and the allocation of planning periods.
[0044] The sequence of the k-th candidate mining body in mining area m is represented as: ; in, This represents the k-th mining body of mining area m in candidate scheme n; k mn Let m be the number of mining bodies included in candidate scheme n.
[0045] The candidate mining body sequence satisfies the mining body precedence relationship; After assigning a planning period configuration τ, a candidate mining plan list c=(σ,τ) is formed. Where: σ is the order of mining bodies; τ is the planning period configuration corresponding to the mining body; Each candidate mining plan meets the following requirements: Each block of ore required to be mined within the corresponding open-pit mine area will be scheduled to be mined once during the entire planning period; Prerequisite relationships between ore blocks and mining bodies; Spatial continuity of the mining body and minimum mining increment requirements; The corresponding single-mine stripping capacity requirements for open-pit mines in each planning period; The pre-set latest completion date requirement for the planning period; Where different mining bodies with a prior relationship are allowed to be configured in the same planning period, the candidate mining plan list also records the order of mining bodies within the same planning period.
[0046] S5) Construct a constrained master problem involving multiple choice set partitioning. set up Open-pit mine The current set of candidate mining plans, Index the candidate mining plans; Configure parameters for the planning period of candidate blocks, when the mining area The The candidate mining plans will list the mineral blocks Arranged during the planning period During mining, ,otherwise . Candidates are listed in the planning period The amount of mining and stripping; This refers to the amount of ore processed; For resources The grade-weighted quantity; For resources Corresponding concentrate yield; This represents the discounted activity cost corresponding to the candidate column. Select variables for candidate columns, when selecting mining areas The When there are 1 candidate column, Otherwise, it is 0. In the linearly relaxed restricted principal problem, In the restricted principal problem of integers, .
[0047] Core attributes such as phased mining and stripping volume, ore input, concentrate output, grade, and discounted cost are extracted from each candidate mining plan. A multi-choice set partitioning constrained master problem is constructed to uniformly constrain global indicators such as shared production capacity, concentrate inventory, order delivery, and blending grade across mining areas, achieving integrated management and control across multiple mining areas. The principle diagram is shown below. Figure 2 .
[0048] The column coefficients are determined based on the phased attributes of the candidate mining plan columns. A multi-choice set is constructed using the candidate column selection variables to divide the restricted master problem. For each mining area, a candidate column for mining blocks with coverage requirements is selected.
[0049] The constrained principal problem of multiple choice set partitioning aims to maximize the project's net present value (NPV) after deducting the discounted operating costs corresponding to the selected candidate mining plans from the actual revenue from concentrate delivery in each planning period, and includes: The candidate column selection variable is used to indicate whether to select the candidate mining plan column for the corresponding open-pit mine area; Candidate column selection constraint is used to select a candidate mining plan column for each open-pit mine that covers all required mining blocks in the open-pit mine and includes arrangements for each planning period; Constraints on single mining area stripping capacity, overall company stripping capacity, and mineral processing capacity are constructed based on the phased coefficients of the candidate mining plan. Concentrate delivery variables and concentrate inventory variables; Concentrate inventory constraints, order delivery constraints, and ore blending constraints are constructed based on the concentrate output and grade-weighted quantities corresponding to the candidate mining plan columns. In the linearly relaxed restricted master problem, the candidate column selection variable is a continuous variable in the interval [0,1], and in the integer restricted master problem, the candidate column selection variable is a binary variable.
[0050] Multiple candidate mining body sequences are generated by changing the seed block, advance direction, weight, or random parameters. After the planning period, a candidate mining plan list is formed, and local feasibility verification is performed.
[0051] Based on the planned mining period of the candidate mining blocks, calculate the stripping volume, ore input, grade-weighted average volume, and concentrate production for each period: ; ; ; ; The discounted operating costs of the candidate mining plan are: ; Construct a multi-choice set partitioning restricted master problem using candidate column selection variables, with the objective function as follows: ; Each mining area selects one column of candidate mining plans that covers the blocks it requires to be mined: ; The phase attributes of the candidate columns are used as column coefficients in the main problem to construct constraints on production capacity, inventory, orders, and ore blending; delivery quantity and inventory quantity are continuous variables, and the column selection variables take [0,1] in linear relaxation and binary values in integer main problems.
[0052] S6) Two-stage collaborative iterative optimization solution; the process is as follows: Figure 3 ; First optimization phase: At least one of the following heuristic optimization algorithms—particle swarm optimization, genetic algorithm, and simulated annealing algorithm—is used to directly generate an initial joint production schedule solution that satisfies the constraints of ore block preconditions, single mining area stripping capacity, cross-mining area sharing capacity, inventory, orders, and ore allocation. The mining body sequence and planning period configuration are extracted from the initial joint production schedule solution according to the mining area to form an initial candidate mining plan column set. Subsequently, the linear relaxed restricted master problem is solved to obtain the dual variables of candidate column selection constraints and cross-mining area constraints.
[0053] The reduced revenue of the candidate mining plan list is expressed as follows: ; in, For the mining area The Reduced returns for each candidate column; The objective coefficient of this candidate column in the objective function of the main problem; For the mining area The dual variable corresponding to the candidate column selection constraint; This is the dual variable vector corresponding to cross-mining area sharing capabilities, inventory, orders, and ore allocation constraints; This is the column coefficient vector of the candidate columns in the above cross-mining area constraints.
[0054] In the formula, the reduced return of the candidate column is determined by its objective contribution, the dual variable of the candidate column selection constraint, the dual variable of the cross-mining area constraint, and the column coefficient.
[0055] Heuristic pricing employs topological expansion, joint sorting of value and dual price, random perturbation, or local exchange to generate new columns; candidate columns with reduced returns and no repetition are added to the column pool, and the integer restricted master problem is solved after the stopping condition is met.
[0056] Second optimization stage: The critical planning period is determined based on the proximity of capacity utilization, inventory, orders, and grade boundaries, and an adaptive large neighborhood search is used to reconstruct the mining body arrangement.
[0057] The adaptive large neighborhood search uses a destruction operator to remove mining bodies and then uses constraint-aware insertion, insertion based on the difference of at least one of the following: marginal change of objective function, change of remaining capacity, change of inventory deviation, change of order deviation, and change of blending grade deviation, or smooth repair and rearrangement between adjacent periods, while updating the production and operation status of each period.
[0058] S7) Candidate column feedback update iteration After a preset number of iterations, the selected weights are updated based on the operator's historical score s, the number of calls n, and the response coefficient ρ. For example...
[0059] in, and Operators Weights before and after the update; For the operator's historical score; For the number of calls; is the reaction coefficient.
[0060] set up For the current solution, For the reconstructed candidate solution, the acceptance probability of the candidate solution is: ; in, The temperature parameter gradually decreases with iteration.
[0061] After partial reconstruction, the changed segments are extracted according to the mining area and recombined with the unmodified parts of the original candidate column to form a feedback candidate mining plan column.
[0062] The feedback columns only check the uniqueness of the mining block, prior relationships, spatial continuity, minimum mining increment, single mining capacity, and latest completion date; global constraints are checked by the main problem combination. Feedback columns that pass the check and are not repeated are added to the candidate column set if they have reducible benefits or originate from a better integer solution.
[0063] 8) Iteration terminates and outputs the final production plan. The two optimization phases are executed alternately. The process stops when any one of the following conditions is met: no new columns are added and the net present value no longer increases, or the maximum time or the maximum number of iterations is reached. The final integer restricted principal problem is then solved, and the final result is the optimal integer feasible solution under the final candidate column set and the preset precision.
[0064] The final output is a long-term plan; The target long-term plan should include at least the following for each open-pit mine in each planning period: mining area, mining body number, ore block number, mining sequence, stripping volume, ore input volume, waste rock volume, concentrate output, initial concentrate inventory, final concentrate inventory, order delivery volume, average grade of ore input, stripping capacity utilization of a single mining area, overall stripping capacity utilization of the company, mineral processing capacity utilization, and project net present value.
[0065] In summary, this invention generates a long-term production plan that satisfies spatial advancement and cross-mining area production and operation constraints by constructing a dynamic mining body, identifying the main problem of candidate plans, generating columns, reconstructing local plans, and updating feedback columns.
[0066] According to the above method, this application also provides an apparatus for generating a long-term production plan scheme for joint open-pit mining in multiple mining areas, characterized in that it includes: The data acquisition module is used to acquire block models, spatial locations, pre-existing relationships of blocks, and production and operation parameters of multiple open-pit mines; The mining body sequence generation module is used to generate mining bodies based on the preceding relationships of ore blocks, spatial adjacency relationships, and minimum mining increments, construct candidate mining body sequences, and configure a candidate mining plan list for the planning period; The set partitioning model construction module is used to establish a reference model for mixed integer programming at the block level, and to construct a multi-choice set partitioning constrained master problem based on the candidate mining plan column; The heuristic column generation module is used to solve the linear relaxed restricted master problem, obtain dual variables, and generate new candidate mining plan columns. An adaptive large neighborhood search module is used to identify critical planning periods and perform destruction and repair on the joint production scheduling solution of multiple mining areas; The sequence splitting feedback module is used to split and reorganize the reconstructed joint production scheduling solution into feedback candidate mining plan columns according to the open-pit mine area and update the candidate column set; The solution output module is used to solve the integer-restricted master problem and output the long-term production plan solution after the preset termination conditions are met.
[0067] It should be understood that those skilled in the art can make improvements or modifications based on the above description, and all such improvements and modifications should fall within the protection scope of the appended claims.
Claims
1. A method for generating a long-term production plan for joint open-pit mining in multiple mining areas, characterized in that, Includes the following steps: 1) Collect production and planning parameters for multiple open-pit mines, including: block models, block attributes, spatial locations, and prior relationships of multiple open-pit mines; 2) Based on the block model, block attributes, and block prerequisites, and combined with economic parameters, establish a block-level mixed integer programming reference model; The reference model for mixed-integer programming at the block level aims to maximize the net present value of a project. The constraints of the model include block preconditions, unique mining constraints, production capacity constraints, inventory constraints, order delivery constraints, and ore blending constraints. 3) For each open-pit mine area, the ore block is divided into multiple mining bodies according to the prior relationship of the ore blocks, the spatial adjacency relationship and the preset minimum mining increment; 4) Generate a sequence of candidate mining bodies based on the pre-existing relationship between the mining bodies, and configure a planning period for each mining body in the candidate mining body sequence to form multiple candidate mining plan columns, wherein each candidate mining plan column contains the mining body sequence and planning period configuration of the mining blocks required to be mined in the corresponding open-pit mining area; 5) Extract the attributes of each candidate mining plan column, including the phased mining and stripping volume, the amount of ore fed into the beneficiation process, the concentrate output, the grade, and the discounted cost, and construct a multi-choice set to divide the constrained master problem in order to select the best single candidate column for each mining area, and uniformly constrain cross-mining area sharing capabilities, concentrate inventory, order delivery, and blending grade. 6) Establish a two-stage collaborative optimization framework to achieve model solution that combines global optimization with local fine-tuning; The first stage is global heuristic optimization; the second stage is local refined reconstruction. 7) Split the reconstructed joint production scheduling solution by mining area and recombin it with the unmodified part of the original candidate column to form a feedback candidate mining plan column; add the feedback column that meets the requirements of local feasibility, no duplication and has positive reduction benefits or comes from a better integer solution to the candidate column set, and return to step 6) for iterative optimization; 8) After the iterative optimization reaches the termination condition, solve the final integer-constrained master problem and output the long-term production plan scheme for joint mining of multiple open-pit mines.
2. The method for generating a long-term production plan for joint open-pit multi-mine mining according to claim 1, characterized in that, In step 2), the objective function of the block-level mixed integer programming reference model is: In the formula, For the planning period set; It is a collection of concentrates or resources; For the planning period The discount factor; For the planning period Internal resources The unit delivery value of the corresponding concentrate; For the planning period Internal resources The actual delivery volume of the corresponding concentrate; For the planning period The costs of mining, transporting, and processing all the mineral blocks within the mine; For initial investment.
3. The method for generating a long-term production plan for joint open-pit multi-mine mining according to claim 1, characterized in that, In step 2), the model construction process includes: determining the concentrate output corresponding to the block based on the block quality, ore or waste rock attributes, resource grade, beneficiation recovery rate, and concentrate grade; determining the block operation cost based on the block quality, mining cost, beneficiation cost, ore transportation cost, and waste rock transportation or disposal cost; determining the concentrate delivery revenue based on the actual concentrate delivery volume and unit delivery value for each planning period; and establishing the block-level mixed integer programming model with the decision variable of whether the block is mined within the corresponding planning period and the objective of maximizing the project's net present value.
4. The method for generating a long-term production plan for joint open-pit multi-mine mining according to claim 1, characterized in that, In step 2), the pre-mining constraint is used to ensure that the pre-mining block corresponding to the ore block scheduled for mining is mined in the same planning period or an earlier planning period. The unique mining constraint of the block is used to ensure that each block is mined at most once during the entire planning period, and to ensure that the blocks required to be mined within the final boundary are mined during the entire planning period. The production capacity constraints include single-mine area mining and stripping capacity constraints, company-wide overall mining and stripping capacity constraints, and mineral processing capacity constraints. The single mining area stripping capacity constraint is used to limit the stripping volume of any open-pit mine in any planning period to not exceed the stripping capacity of that open-pit mine. The company's overall mining and stripping capacity constraint is used to limit the total mining and stripping volume of all open-pit mines within the same planning period from not exceeding the company's overall mining and stripping capacity. The mineral processing capacity constraint is used to limit the total amount of ore fed into all open-pit mines within the same planning period to no more than the mineral processing capacity.
5. The method for generating a long-term production plan for joint open-pit multi-mine mining according to claim 1, characterized in that, In step 3), the method for dividing the mining body includes: The set of currently mineable blocks is determined based on the blocks that have not yet been included in the mining body and their preceding relationships. Select a seed block from the current set of mineable blocks. Blocks that are spatially adjacent to the current set of blocks and whose predecessor blocks are already located in the previous mining body or the current set of blocks are designated as expandable blocks. The expansion order is determined according to economic value, spatial distance, direction of advance, or resource grade until the minimum mining increment is reached.
6. The method for generating a long-term production plan for joint open-pit multi-mine mining according to claim 1, characterized in that, In step 5), the column coefficients are determined based on the phased attributes of the candidate mining plan columns, and the candidate column selection variables are used to construct a multi-selection set to divide the restricted principal problem.
7. The method for generating a long-term production plan for joint open-pit mining in multiple mining areas according to claim 6, characterized in that, In step 5), the column coefficients are determined based on the phased attributes of the candidate mining plan columns, and the candidate column selection variables are used to construct a multi-selection set to divide the restricted principal problem.
8. The method for generating a long-term production plan for joint open-pit mining in multiple mining areas according to claim 1, characterized in that, In step 6), Phase 1: Generate initial joint production scheduling solutions and initial candidate column sets that satisfy full constraints for single mining area and cross mining area; solve the linear relaxation master problem to obtain dual information, use the reduction benefit as a screening index, continuously iterate to generate high-quality new candidate columns, and after the iteration reaches the standard, solve the integer-constrained master problem to obtain the basic optimal solution; The second stage involves calculating the capacity utilization rate and the proximity of inventory to grade boundaries for each planning period to screen key planning periods. An adaptive large neighborhood search algorithm is then used to reconstruct the mining arrangement for key periods by using destruction and repair operators. The weights are dynamically updated based on the historical optimization effects of the operators to achieve refined optimization of local solutions.
9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 8.
10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method described in any one of claims 1 to 8.