Engineering constraint coupled stacker stereoscopic warehouse layout generation method

Abstract

The application discloses an engineering constraint coupled stacker stereoscopic warehouse layout generation method, and relates to intelligent warehouse layout optimization and computer aided design. The method collects factory building, material, goods shelf, stacker and project demand parameters, and establishes an installability rule library; parameterized modeling is performed on single-deep and double-deep goods shelf cell units to generate installable templates; two-layer collaborative optimization is used to solve the engineering installable template set, the number of roadways, and the stacker speed parameter combination, and combined with local installation conflict identification, illegal solution repair and reverse correction, a Pareto optimal solution set is obtained; an installation level parameter package is formed and output after consistency verification, and a layout drawing and a parameter list are output. The method can consider cost, space utilization and operation efficiency, and improve the engineering installability of the layout scheme.

Description

Technical Field

[0001] This invention relates to the field of intelligent warehouse layout optimization and computer-aided design technology, and in particular to an engineering layout generation method for stacker crane automated storage and retrieval systems. Background Technology

[0002] In the field of intelligent warehousing and logistics automation, stacker crane automated storage and retrieval systems (AS / RS) have become crucial infrastructure for manufacturing, commercial logistics, and supply chain centers due to their advantages such as high space utilization, high automation, and high operational efficiency. Extensive research has been conducted both domestically and internationally on the design, layout, and performance evaluation of AS / RS. For example, the 2010 paper "Research on warehouse design and performance evaluation: A comprehensive review" published in the *European Journal of Operational Research* systematically reviewed warehouse design, performance evaluation, and computer-aided tools, demonstrating that warehouse design has become an important direction in warehousing system research.

[0003] In existing technologies, one type of approach focuses on the generation and optimization of warehouse spatial layout. For example, patent document CN117875189B, "A Method for Spatial Layout of Automated Warehouse Based on GA Optimization GRO," discloses a technical solution that first uses the SLP method to generate an initial layout diagram of the automated warehouse, then establishes a digital model, and uses the GA optimization GRO algorithm to optimize the positions of functional units. The paper "A novel framework for automated warehouse layout generation" published by Frontiers in Artificial Intelligence in 2024 also proposes an automatic layout generation framework that generates warehouse layouts under given spatial parameters and verifies their feasibility. These approaches demonstrate that using intelligent algorithms or constraint search methods to generate and optimize warehouse layouts has become an important technical approach in current research.

[0004] Another approach focuses on multi-objective optimization and parameter design in automated storage and retrieval systems (AS / RS). For example, patent document CN115303689A, "An Optimization Method for Storage Location Allocation in a Multi-Aisle AS / RS," comprehensively considers factors such as inbound / outbound efficiency, rack stability, and stacker crane load balancing to construct a multi-objective optimization model and uses an improved adaptive multi-population genetic algorithm for solution. Patent document CN114417696B, "An Optimization Method for Storage Location Allocation in an Automated Storage and Retrieval System Based on a Genetic Algorithm," comprehensively considers operation time, rack stability, and product correlation, using an improved genetic algorithm for optimization. The paper "Automated storage and retrieval system design with variant lane depths," published in the European Journal of Operational Research in 2024, further models AS / RS design as an optimization problem that balances investment and throughput capacity, incorporating the number of racks, rack dimensions, and equipment parameters into the design decisions. Therefore, conducting multi-objective modeling and optimization solutions for automated storage and retrieval systems (AS / RS) is a mature research direction in existing technologies.

[0005] However, existing technologies still have at least the following shortcomings: First, the collaborative processing of multiple objectives is still insufficient. Although some existing solutions have introduced multi-objective models, they mostly focus on location allocation, operation scheduling, or optimization of local functional areas, or mainly analyze investment costs and throughput capacity. They are still difficult to conduct unified, systematic, and engineering-implementation-oriented collaborative optimization of construction costs, space utilization, and operational efficiency in the overall layout design stage of stacker crane automated warehouses. Therefore, they cannot fully meet the multidimensional needs of enterprises in actual project decision-making.

[0006] Secondly, there is insufficient consideration of the physical constraints of actual engineering installation. For example, existing research on automatic layout generation typically focuses on the arrangement of functional areas, aisle connectivity, minimum clearance, or accessibility verification within a given spatial range; existing patents on automated warehouse optimization focus more on location allocation, operational efficiency, or optimization of functional unit positions. For installation-level physical constraints commonly found in stacker crane automated warehouse engineering design, such as stacker crane end dimensions, front and rear safety spaces, rack column hole spacing, beam thickness, height differences between top and non-top racks, structural differences in different depths of double-deep racks, and fire protection reserved space, existing solutions often lack system coupling modeling. This can easily lead to situations where, although the generated solution is mathematically feasible, it may be impossible to install, have an unreasonable layout, or insufficient safety boundaries in actual engineering.

[0007] Third, existing algorithms still lack adaptability to mixed decision variables. Stacker crane automated warehouse layout design typically involves integer, continuous, and discrete decision variables simultaneously. For example, the number of aisles is usually an integer, and parameters such as stacker crane speed and acceleration often need to be selected from discrete values ​​given in equipment samples or product catalogs. While existing genetic algorithms or improved genetic algorithm schemes can be used for optimization solutions, they are mostly designed for general encoding and crossover / mutation processes, lacking specialized encoding, constraint filtering, and illegal solution repair mechanisms for mixed "integer-discrete" variables. This easily leads to a large number of illegal solutions, affecting convergence efficiency and engineering applicability.

[0008] Therefore, existing technologies still need to address the following technical issues: how to provide a method for designing the layout of automated warehouses oriented towards stacker cranes, which can collaboratively optimize warehouse construction costs, space utilization, and operational efficiency while meeting the hard constraints of engineering installation such as stacker crane end dimensions, rack column hole spacing, beam thickness, fire protection provisions, and safety clearance, and improve the effectiveness of solving mixed integer-discrete decision variables, thereby generating automated warehouse layout schemes that can directly serve actual engineering design and installation. Summary of the Invention

[0009] To address the challenges in existing stacker crane automated warehouse layout designs, such as the difficulty in simultaneously optimizing construction costs, space utilization, and operational efficiency, the lack of systematic coupling of actual engineering installation constraints (e.g., stacker crane end buffer space, rack column hole spacing, beam thickness, top and bottom safety space of racks, fire protection reserved space), and the tendency to generate numerous illegal solutions, low solution efficiency, and difficulty in directly outputting layout results usable for engineering installation when optimizing mixed decision variables such as the number of aisles and stacker crane speed parameters, this invention proposes a stacker crane automated warehouse layout generation method with coupled engineering constraints. This method achieves automatic generation and engineering implementation of stacker crane automated warehouse layout schemes through pre-definition of basic engineering parameters and constraints, parametric modeling of storage cell units coupled with engineering constraints, solving for feasible solution sets of single-row racks, multi-objective collaborative optimization of aisle number and stacker crane speed parameters, and engineering layout output.

[0010] The technical solution adopted in this invention is as follows: A method for generating the layout of a stacker crane automated warehouse with engineering constraints coupling includes the following steps: S1. Collect the available space parameters of the factory, material unit parameters, rack parameters, stacker crane parameters and project requirement parameters required for the design of the stacker crane automated warehouse, and establish an installability rule library based on the collected parameters; S2. Based on the parameters and installability rule base in step S1, perform parametric modeling of single-deep racking unit and double-deep racking unit, and generate corresponding installable templates according to different depths, different levels and different aisle locations. S3. With the goal of minimizing space waste in the length and height directions of the factory building, and with constraints such as the stacker crane end buffer space, the safety space at the top and bottom of the rack, the matching of the upright hole spacing, the fire protection reserved space and the factory boundary, the first-level optimization is carried out on single-deep racks and double-deep racks respectively, the number of columns and layers of single-row racks is solved, and a set of engineering installable templates with installation rule labels and reverse correction interfaces is generated. S4. Using only the set of installable templates generated in step S3 as the solution basis, and with multiple objectives of minimizing space waste in the width direction of the factory, minimizing overall cost, and maximizing operational efficiency, a second layer of optimization is performed on the combination of single-depth aisle number, double-depth aisle number, semi-double-depth aisle number, and stacker crane speed parameters. During the iterative optimization process, local installation conflict identification is performed on the generated candidate solutions. When there is a repairable conflict, illegal solution repair is performed according to the installability rule base. When there is an installation conflict that cannot be directly repaired in the current layer, a reverse correction mechanism is triggered, and the feasible template parameters in step S3 are backtracked and corrected before re-optimization and evaluation, finally obtaining the Pareto optimal solution set. S5. Select target solutions from the Pareto optimal solution set in step S4, and extract layout parameters, equipment selection parameters, installation hole parameters, clearance parameters, safety boundary parameters and bill of materials parameters to form an installation-level parameter package; S6. Perform a consistency check on the installation-level parameter package in step S5. After the check passes, drive the drawing generation system to output the layout drawings and parameter list of the automated warehouse that can be directly used for engineering installation.

[0011] Furthermore, the installability rule base established in step S1 includes at least: a geometric installation rule sub-base, a layer-level differentiation rule sub-base, a depth-level differentiation rule sub-base, a location-level differentiation rule sub-base, a reverse correction rule sub-base, and a delivery consistency verification rule sub-base. The geometric installation rule sub-library constrains length, width, height, end buffer space, aisle clear width, back-pull space, and back-to-back spacing; the layer differentiation rule sub-library distinguishes the height and safety space requirements of top-level and non-top-level storage cells; the depth differentiation rule sub-library distinguishes the size and operational boundary requirements of single-deep storage cells, double-deep first-deep storage cells, double-deep second-deep storage cells, and semi-double-deep scenarios; the location differentiation rule sub-library distinguishes the installation templates for end aisles, middle aisles, side shelves, and back-to-back shelves; the reverse correction rule sub-library determines the backtracking object, correction order, and correction scope when installation conflicts are found in the second-level optimization; and the delivery consistency verification rule sub-library performs cross-checking of layout parameters, equipment selection parameters, installation hole parameters, clearance parameters, safety boundary parameters, and bill of materials parameters.

[0012] Furthermore, in step S2, the parametric modeling of single-deep racking units and double-deep racking units includes at least: single-deep racking units and double-deep racking units; The width of a single-depth shelving unit is: ; The length of a single-depth shelving unit is: ; The height of non-top-level shelves in a single-depth rack is: ; The height of the top shelf compartment in a single-depth shelving unit is: ; The heights of the first and second deep non-top shelf compartments in double-deep shelving are as follows: ; The height of the top shelf compartment in the deepest position of a double-deep shelving unit is: ; The height of the top shelf compartment in the second deep position of a double-deep shelving unit is: ; The height of non-top-level shelves is calculated by rounding up and taking into account the center distance between the shelf upright holes. Alignment is performed to ensure that the height of the storage compartment matches the mounting holes on the shelf; In the formula, Indicates the width of the storage compartment; Indicates the length of the storage compartment; Indicates the height of non-top-level shelves in a single-depth shelving unit; Indicates the height of the top shelf compartment in a single-depth shelving unit; This indicates the height of the first, non-top-level shelf compartment in a double-deep shelving unit. This indicates the height of the second deep shelf unit (excluding the top shelf) in a double-deep shelving system. This indicates the height of the top shelf compartment in the deepest position of a double-deep shelving unit. This indicates the height of the top shelf compartment in the second deepest position of a double-deep shelving unit; , , These represent the width, length, and height of the stored material, respectively. Indicates the number of storage locations in a single storage compartment; Indicates the spacing between materials; Indicates the distance between the materials and the shelf uprights; Indicates the width of the shelf uprights; This indicates the safe distance that the material should extend beyond the crossbeam of the storage compartment; This indicates raising the safety space; Indicates space reserved for fire protection; Indicates the height of the beam; Indicates the center distance between the holes in the shelf uprights; Indicates the safety space at the top of a single-depth shelving unit; Indicates the safety space at the top of double-deep shelving; This indicates the additional lifting space required for double-deep shelving; This indicates the rounding up operation.

[0013] Furthermore, in step S3, the first layer of optimization establishes target programming models for single-deep and double-deep shelves respectively, with the number of shelf columns and shelf layers as decision variables, and all decision variables are positive integers; The optimization model for the first level of a single-depth rack, aiming to minimize space waste in both the length and height directions of the factory building, is expressed as follows: ; The constraints are: ; ; ; ; , Represents the set of positive integers; The optimization model for the first level of double-deep shelving, aiming to minimize space waste in both the length and height directions of the factory building, is expressed as follows: ; The constraints are: ; ; ; ; ; After solving the first layer of optimization, the single-deep rack template and double-deep rack template that satisfy the constraints are represented as follows: ; ; in, This represents the objective function value for optimizing the first layer of a single-depth shelving unit. This represents the objective function value for optimizing the first layer of a double-deep shelving unit. and These represent the total length and total height occupied by a single-depth rack or a single row of racks, respectively. and These represent the total length and total height occupied by a single row of double-depth shelving units, respectively. and These represent the length and height of the factory building that can be used for stacker crane automated warehouse layout; This indicates the length of space occupied by one side of the stacker crane's end buffer. Indicates the height of the safety space at the bottom of the shelf; Indicates the width of the storage compartment; Indicates the height of the top shelf compartment in a single-depth shelving unit; Indicates the height of non-top-level shelves in a single-depth shelving unit; This indicates the height of the top shelf compartment in the deepest position of a double-deep shelving unit. This indicates the height of the first, non-top-level shelf compartment in a double-deep shelving unit. , , , These represent the number of rows, layers, and levels of a single-deep shelving unit, respectively, and are all positive integers. and These represent the sets of templates that can be installed in single-depth racking projects and double-depth racking projects, respectively. , This represents the optimal number of columns and optimal number of layers for a single-depth shelving unit obtained from the first level of optimization. , This represents the optimal number of columns and optimal number of layers for the double-deep shelving obtained from the first level of optimization; , This represents the total occupied length and total occupied height corresponding to the optimal single-depth rack solution; , This indicates the total occupied length and total occupied height corresponding to the optimal solution for double-depth shelving; , Indicates the installation rule label; , Indicates a template index; , This indicates the reverse correction interface.

[0014] Furthermore, in step S4, the decision variables for the second-level optimization include integer decision variables and discrete decision variables. The discrete decision variables include the combination of single-depth stacker crane speed parameters. Combined speed parameters with double-deep stacker crane ,in: ; ; ; ; The objective function of the second-level optimization should include at least the following: minimizing space waste in the width direction of the plant, minimizing the comprehensive cost consisting of stacker crane cost and storage location cost, and maximizing the total operating efficiency of stacker cranes in all aisles. Minimizing space waste in the width direction of the factory building is expressed as: ; The minimum overall cost is expressed as: ; The maximum overall operating efficiency of all stacker cranes in all aisles is expressed as: ; The constraints include: number of storage locations, operational efficiency, aisle width, plant width boundary, and semi-double-deep stacker crane efficiency. The quantity constraint for storage locations is: ; ; The work efficiency constraint is: ; The width constraint of the tunnel is: ; ; ; The factory building width boundary constraint is: ; The efficiency constraint for a semi-double-deep stacker is: ; In the formula, These represent the number of aisles for single-deep stacker cranes, double-deep stacker cranes, and semi-double-deep stacker cranes, respectively. Represents the set of non-negative integers; This indicates the combination of speed parameters for a single-depth stacker crane. This indicates the combination of speed parameters for a double-deep stacker crane; This represents a preset discrete set of values ​​for the speed parameter combinations of a single-depth stacker crane. This represents a preset discrete set of values ​​for the speed parameter combinations of a double-deep stacker crane; This indicates the travel speed of a single-depth stacker crane; This indicates the lifting speed of a single-depth stacker crane; This indicates the speed of the telescopic fork of a single-depth stacker crane; This indicates the travel speed of the double-deep stacker; This indicates the lifting speed of the double-deep stacker; Indicates the speed of the telescopic fork of a double-deep stacker; This represents the objective function value indicating the waste of space in the width direction of the factory building; This represents the value of the overall cost objective function; This represents the objective function value for the overall operating efficiency of stacker cranes in all aisles; This indicates the width of the factory building that can be used for stacker crane automated warehouse layout; Indicates the width of a single-depth tunnel; Indicates the width of a double-deep tunnel; Indicates the width of a semi-double-deep tunnel; This indicates the back-to-back spacing between shelves in adjacent aisles; Indicates the space required for back-pull shelving; This indicates the speed parameter combination of a single-depth stacker crane. The corresponding equipment cost; This indicates the speed parameter combination of the double-deep stacker crane. The corresponding equipment cost; This indicates the cost per storage location; This indicates the total number of storage locations in the plan; This indicates the speed parameter combination of a single-depth stacker crane. Total length occupied by a single row of shelves Total height occupied by single-row shelving Operational efficiency under certain conditions; This indicates the speed parameter combination of the double-deep stacker crane. Total length occupied by a single row of shelves Total height occupied by single-row shelving Operational efficiency under certain conditions; This indicates the operational efficiency corresponding to a semi-double-depth tunnel. Indicates the efficiency of project requirements; Indicates the number of storage locations in a single storage compartment; This represents the optimal number of columns for a single-depth shelving unit obtained from the first level of optimization. This represents the optimal number of single-depth shelves obtained from the first level of optimization; This represents the optimal number of columns for the double-deep shelving obtained from the first level of optimization; This represents the optimal number of layers for the double-deep shelving obtained from the first layer of optimization; Indicates the minimum number of storage locations required for the project; This indicates the total number of cargo units corresponding to a single-depth aisle; This indicates the total number of cargo units corresponding to a double-deep aisle; This indicates the total number of cargo units corresponding to a semi-double-depth aisle; This indicates the distance between materials on the loading platform and materials on the shelves in the aisle; Indicates the length of a single stored material; This indicates the distance between the first and second deep items in a double-deep shelving unit; This indicates the total length occupied by a single-depth rack or a single row of racks; This indicates the total height occupied by a single-depth rack or a single row of racks; This indicates the total length occupied by a single row of double-deep shelving; This indicates the total height occupied by a single row of double-deep shelving.

[0015] Furthermore, step S4 employs an improved genetic algorithm with adapted mixed decision variables for solution, the improved genetic algorithm comprising: The number of tunnels is encoded using integers; Discrete index encoding is used for the stacker crane speed parameter combinations; The feasible templates generated in step S3 are encoded using template indexing. During fitness evaluation, both multi-objective function values ​​and constraint violation rates are calculated simultaneously. Individuals are screened using a non-dominated ranking system prioritizing constraint dominance and crowding calculation. Tournament selection was used in the selection process; During the crossover process, integer-coded genes are exchanged, discrete-index-coded genes are exchanged within the directory index range, and template-index-coded genes are exchanged within the same rule category. During the mutation process, random integer mutations are performed on integer-coded genes, random index mutations within the directory are performed on discrete-index-coded genes, and template-index-coded genes are subjected to similar template-neighborhood mutations.

[0016] Furthermore, the local installation conflict identification in step S4 includes at least one or more of the following: Conflicts include: end interference, insufficient clearance, misalignment of installation holes, mismatch of rules between top and non-top levels, mismatch of rules between first and second depth levels, inconsistency of equipment operating boundaries, and mismatch of position templates. When a minor conflict is detected, the current layer is repaired based on the reverse correction rule sub-library for the number of lanes, template index, or velocity parameter index. When a moderate to severe conflict is identified, the reverse correction rule sub-library is used to backtrack to step S3 to correct the number of shelf layers, shelf columns, top-level template, deep-level template, or location template, and then the corrected template is sent back to step S4 for optimization evaluation.

[0017] Furthermore, the installable templates generated in step S2 include at least: single-depth non-top-level template, single-depth top-level template, double-depth first-depth non-top-level template, double-depth second-depth non-top-level template, double-depth first-depth top-level template, double-depth second-depth top-level template, end aisle template, middle aisle template, side shelf template, back-to-back shelf template, and semi-double-depth template; Each template is associated with the corresponding installation rule label, clearance boundary parameters, installation hole boundary parameters, and equipment operation boundary parameters.

[0018] Furthermore, the installation-level parameter package formed in step S5 includes at least: storage unit size parameters, rack layer and column number parameters, aisle centerline and aisle width parameters, stacker crane model and speed parameters, end buffer zone boundary parameters, back-to-back spacing parameters, back-pull space parameters, mounting hole sequence parameters, top clearance parameters, bottom clearance parameters, fire clearance parameters, and bill of materials parameters.

[0019] Furthermore, the consistency verification in step S6 includes at least the following: consistency verification between layout parameters and equipment selection parameters; consistency verification between rack height parameters and installation hole parameters; consistency verification between aisle width parameters and stacker crane operating boundary parameters; consistency verification between clearance parameters and safety boundary parameters; consistency verification between drawing output parameters and bill of materials parameters; and consistency verification between installation rule labels and final delivery parameters. After all consistency checks pass, the drawing generation system outputs the warehouse unit diagram, automated warehouse floor plan, automated warehouse elevation layout, stacker crane parameter table, installation hole location table, clearance check table, and equipment bill of materials table.

[0020] Compared with existing technologies, this invention couples the hard constraints of engineering installation to the entire layout generation process in advance, and combines a two-layer collaborative optimization, local conflict repair and reverse correction mechanism, and installation-level parameter package consistency verification to achieve installable generation, collaborative optimization solution, and engineering delivery output of stacker crane automated warehouse layout schemes. Therefore, it has at least the following beneficial effects: 1. This invention achieves layout generation oriented towards actual engineering installation constraints, significantly reducing the probability of uninstallable solutions: Before layout generation, this invention collects available space parameters, material unit parameters, rack parameters, stacker crane parameters, and project requirement parameters from the factory floor, and establishes an installability rule library. It pre-embeds hard engineering installation constraints such as stacker crane end buffer space, rack top and bottom safety space, column hole spacing, fire protection reserved space, aisle clear width, back-pull space, and back-to-back spacing into the entire layout generation process. Simultaneously, during the parametric modeling stage of the storage unit, the height of non-top-level storage units is aligned with the center distance of the column holes, and corresponding installable templates are generated for different depths, levels, and aisle locations. This ensures that the generated results possess engineering installability from the outset, avoiding the problems of existing technologies that are only mathematically feasible but cannot be installed in actual engineering, have unreasonable layouts, or insufficient safety boundaries.

[0021] 2. This invention achieves synergistic optimization of construction costs, space utilization, and operational efficiency, which is conducive to improving the overall decision-making quality of the overall layout scheme: This invention does not only optimize for a single objective, but also simultaneously aims at minimizing space waste in the width direction of the factory building, minimizing overall costs, and maximizing the total operational efficiency of stacker cranes in all aisles in the second-level optimization. It performs multi-objective synergistic optimization on the number of single-depth aisles, double-depth aisles, semi-double-depth aisles, and stacker crane speed parameters. Therefore, it can comprehensively balance the relationship between warehouse construction costs, space utilization, and operational efficiency while meeting the project's required number of storage locations and project required efficiency. Compared with existing schemes that only focus on the optimization of local functional areas, storage location allocation, or single cost / efficiency indicators, this invention is more suitable for the multi-dimensional decision-making needs of enterprises in actual engineering projects.

[0022] 3. The "two-layer optimization" architecture enhances the solution's specificity and engineering adaptability, helping to reduce the inefficiency caused by full-range searches: This invention first uses a first-layer optimization to solve for the set of installable templates for single-row engineering of single-depth and double-depth racks, respectively. Then, it uses only this set of installable templates as the basis for the second-layer optimization. In other words, the second-layer optimization does not blindly search the entire parameter space, but rather further optimizes within the range of feasible templates that already meet constraints such as length, height, and installation boundaries. This effectively reduces the search space size, decreases the number of candidate solutions that do not meet the engineering boundary conditions from participating in the optimization calculation, and thus improves the overall solution efficiency and the validity of the results.

[0023] 4. Improved adaptability to mixed integer-discrete decision variables, reducing the number of illegal solutions and improving optimization convergence: This invention addresses the engineering characteristics of stacker crane automated warehouse layout design, where the number of aisles is an integer variable and the stacker crane speed parameters are discrete variables from the equipment catalog. In the second-level optimization, an improved genetic algorithm adapted to the mixed decision variables is employed: the number of aisles is encoded with integers, the stacker crane speed parameter combination is encoded with discrete indexes, and the template is encoded with template indexes. During crossover and mutation, these variations are limited to the catalog index range or the same rule category. Compared to general encoding and general crossover / mutation strategies, this invention reduces the generation of illegal solutions from the encoding source and improves the algorithm's adaptability to actual equipment selection and engineering parameter constraints during multi-objective optimization.

[0024] 5. By employing a mechanism for identifying local installation conflicts, repairing illegal solutions, and implementing reverse correction, the utilization rate of candidate solutions and the feasibility of the final result are improved. During the second-level optimization iteration, this invention performs local installation conflict identification on candidate solutions. The identified conflicts include at least end interference conflicts, insufficient clearance conflicts, misaligned mounting hole conflicts, top-level and non-top-level rule mismatch conflicts, first-depth and second-depth rule mismatch conflicts, inconsistent equipment operating boundaries, and position template mismatch conflicts. For repairable conflicts, the rule base is used to repair illegal solutions at the current level. For conflicts that cannot be directly repaired at the current level, a reverse correction mechanism is triggered, backtracking and correcting the feasible template parameters in the first level, and then re-evaluating the optimization. This not only avoids the information waste caused by directly discarding candidate solutions with local installation conflicts in traditional methods, but also makes the final Pareto optimal solution set closer to the feasible results of actual engineering projects.

[0025] 6. This invention achieves direct, interconnected output from optimization results to installation-level delivery results, reducing manual conversion costs and improving project delivery consistency: After obtaining the Pareto optimal solution set, this invention further extracts layout parameters, equipment selection parameters, mounting hole parameters, clearance parameters, safety boundary parameters, and bill of materials parameters to form an installation-level parameter package, and performs consistency verification on it. Only after the consistency verification passes is the drawing generation system driven to output the storage unit diagram, automated warehouse floor plan, elevation layout, stacker crane parameter table, mounting hole table, clearance verification table, and equipment bill of materials table. Therefore, this invention achieves integrated linkage between layout design, equipment selection, installation boundary control, and engineering drawing output, reducing the error risks caused by manual secondary processing, manual conversion, and manual verification in the traditional design process, and improving the accuracy and consistency of project delivery.

[0026] 7. This invention is particularly suitable for the design of stacker crane automated storage and retrieval systems (AS / RS) with complex installation boundaries and multiple types of aisle combinations, and has good engineering practical value: Because this invention considers single-depth, double-depth, and semi-double-depth aisle combination scenarios, and establishes differentiated installable templates for different depths, different levels, and different aisle locations, and integrates equipment selection parameters, storage location quantity constraints, efficiency constraints, and installation-level output requirements into the same technology chain, it can better adapt to the layout design needs of AS / RS in complex engineering scenarios, and has strong engineering feasibility and promotion application value. Attached Figure Description

[0027] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0028] Figure 1 This is the overall flowchart of the stacker crane automated warehouse layout generation method with engineering constraint coupling of the present invention; Figure 2 , Figure 3 This is a flowchart of parametric modeling and template generation for the S2 coupled engineering constraints of the present invention. Figure 4 This is a schematic diagram of the main view of a non-top-level storage unit of the present invention; Figure 5 This is a schematic diagram of the top-level storage unit of the present invention; Figure 6 This is a side view of the cargo compartment unit of the present invention; Figure 7 This is a flowchart of the first-layer engineering installability generation mechanism based on constraint propagation and pruning rules in S3 of the present invention; Figure 8 This is a flowchart of the second-layer dedicated solution mechanism for the S4 of the present invention, which is oriented towards the discrete values ​​of the device catalog and the repair projection of illegal solutions. Figure 9 This is a CAD layout drawing of an automated warehouse output by the parametric drawing generation system of this invention. Detailed Implementation

[0029] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0030] This embodiment provides a method for generating layouts of stacker crane automated warehouses with coupled engineering constraints. This method addresses the engineering design, equipment selection, installation verification, and drawing delivery scenarios of stacker crane automated warehouses. It revolves around a technical chain of "pre-engineering constraint generation - storage compartment template generation - two-layer collaborative optimization - local conflict repair and reverse correction - installation-level parameter package output." By embedding hard engineering installation constraints, discrete equipment selection constraints, and construction delivery constraints into a unified process, it achieves closed-loop processing from candidate layout generation to installation-level drawing output.

[0031] like Figure 1 As shown, in this embodiment, the method for generating the layout of a stacker crane automated warehouse with engineering constraints coupling includes the following steps: S1. Engineering basic parameter acquisition, constraint pre-definition, and installationability rule base establishment: The function of this step is to collect all the basic parameters required for the engineering design of the stacker crane automated warehouse, and to make explicit and structured the implicit constraints that were originally scattered in building conditions, racking installation conditions, stacker crane sample catalogs, fire protection boundaries, and construction drawing review, forming a common engineering constraint basis for subsequent steps. This step is not simply about entering parameters, but rather "engineering constraints first".

[0032] In this embodiment, the collected parameters of available factory space include at least the available length of the factory building. Available width of factory building Available height of factory buildings The material unit parameters should at least include the length of the stored material. ,width and height The racking parameters should at least include the width of the racking uprights. Center distance of column holes Beam height Number of storage locations per unit Spacing between materials Spacing between materials and columns The safe distance of materials exceeding the crossbeams of the storage compartment Raise the safety space Fire safety reserved space Safety space at the bottom of the shelf Single-depth top safety space Double-deep top safety space Double-deep additional lifting space The distance between materials on the loading platform and materials on the shelves in the aisle The distance between the first and second depth positions of the double-deep system Back-to-back spacing and backrest space The stacker crane parameters should at least include the combination of speed parameters for a single-depth stacker crane. Speed ​​parameter combination of double-deep stacker crane End buffer space Price of a single-depth stacker crane Price corresponding to double-deep stacker crane The project requirements parameters should at least include the required number of storage locations. and project requirements efficiency .

[0033] In this embodiment, the installability rule base is further subdivided into the following rule modules: Firstly, the geometric installation rules module is used to constrain the length, width, and height boundaries of the plant, the stacker crane end buffer zone, back-pull space, back-to-back spacing, aisle net width, minimum offset of track centerline, and equipment operating envelope space. Secondly, the layer differentiation rule module is used to distinguish the different requirements of top-level and non-top-level storage compartments in terms of height, top safety space, and lifting space for telescopic forks. Third, the depth differentiation rule module is used to distinguish the differences in storage and retrieval boundaries, effective operating depth of telescopic forks, and safety distance in single-deep, double-deep first-deep, double-deep second-deep, and semi-double-deep scenarios. Fourth, the location differentiation rule module is used to distinguish the template boundaries corresponding to end aisles, middle aisles, side shelves, back-to-back shelves, and semi-double-depth aisles. Fifth, the equipment catalog constraint module is used to stipulate that the combination of speed parameters for single-deep and double-deep stacker cranes can only be selected from the discrete values ​​given in the actual equipment catalog, and to record the equipment model, cost, adaptable depth, allowable service height, allowable service length and minimum aisle width corresponding to each set of speed parameters. Sixth, the reverse correction rule module is used to determine the priority backtracking template, parameter correction order, allowable correction step size and stopping condition when conflicts occur in the second-level optimization. Seventh, the delivery consistency verification module is used to perform item-by-item cross-checking of layout parameters, equipment parameters, hole position parameters, clearance boundary parameters and BOM parameters after the installation-level parameter package is formed.

[0034] Through the above settings, this step not only defines the engineering parameters, but also transforms the "engineering installation rules" into a data structure that can be called, propagated, pruned, repaired, and delivered for verification in subsequent calculations, thereby providing a constraint basis for the subsequent engineering installability generation mechanism.

[0035] S2. Parametric modeling and template generation of cell elements with coupled engineering constraints: The function of this step is to: use the engineering parameters and rule base collected in step S1 as input to establish standard storage cell units for single-deep and double-deep racks, and further generate installable templates with differentiated installation boundaries. The purpose of this step is not simply to calculate a few dimensions, but to upgrade the "storage cell geometry" to a "structural template with engineering constraint labels," providing a foundation for the subsequent generation and pruning of the first-level engineering installability.

[0036] like Figure 2 , Figure 3 As shown, in this embodiment, the parametric modeling of single-deep and double-deep racking units includes at least the following formulas: 1. For single-depth racking unit: The width of a single-depth shelving unit is: ; The length of a single-depth shelving unit is: ; The height of non-top-level shelves in a single-depth rack is: ; The height of the top shelf compartment in a single-depth shelving unit is: ; In the formula, Indicates the width of the storage compartment; Indicates the length of the storage compartment; Indicates the height of non-top-level shelves in a single-depth shelving unit; Indicates the height of the top shelf compartment in a single-depth shelving unit; , , These represent the width, length, and height of the stored material, respectively. Indicates the number of storage locations in a single storage compartment; Indicates the spacing between materials; Indicates the distance between the materials and the shelf uprights; Indicates the width of the shelf uprights; This indicates the safe distance that the material should extend beyond the crossbeam of the storage compartment; This indicates raising the safety space; Indicates space reserved for fire protection; Indicates the height of the beam; Indicates the center distance between the holes in the shelf uprights; Indicates the safety space at the top of a single-depth shelving unit; This indicates the rounding up operation.

[0037] 2. For double-deep shelving unit: The heights of the first and second deep non-top shelf compartments in double-deep shelving are as follows: ; The height of the top shelf compartment in the deepest position of a double-deep shelving unit is: ; The height of the top shelf compartment in the second deep position of a double-deep shelving unit is: ; The height of non-top-level shelves is calculated by rounding up and taking into account the center distance between the shelf upright holes. Align the shelves to ensure that the shelf height matches the mounting holes on the rack.

[0038] In the formula, This indicates the height of the first, non-top-level shelf compartment in a double-deep shelving unit. This indicates the height of the second deep shelf unit (excluding the top shelf) in a double-deep shelving system. This indicates the height of the top shelf compartment in the deepest position of a double-deep shelving unit. This indicates the height of the top shelf compartment in the second deepest position of a double-deep shelving unit; Indicates the safety space at the top of double-deep shelving; This indicates the additional lifting space required for double-deep shelving.

[0039] The correspondence between the single-depth / double-depth racking unit parameters involved in steps S1 and S2 is as follows: Figures 4 to 6 As shown, where, Figures 4 to 6 The diagram shows the corresponding positional relationships of parameters such as storage compartment width, storage compartment height, material height, material width, material length, distance between materials, distance between materials and uprights, safe distance of materials exceeding the storage compartment beams, upright width, beam height, lifting safety space, space required for double-deep racks, top safety space, and fire protection reserved space in single-deep or double-deep storage compartment units. This facilitates understanding the geometric basis and engineering constraint sources for parametric modeling of storage compartment units in step S2.

[0040] After obtaining the basic unit dimensions, this embodiment further generates tagged templates instead of directly feeding these dimensions into the optimization process. Specifically, template generation includes at least the following processes: First, templates are used to break down the differences in shelf positions. For single-depth shelving, separate templates are generated for single-depth non-top shelf positions and single-depth top shelf positions. For double-depth shelving, separate templates are generated for double-depth first-depth non-top shelf positions, double-depth second-depth non-top shelf positions, double-depth first-depth top shelf positions, and double-depth second-depth top shelf positions. In this way, the different clearance requirements between the top and non-top shelves, the different elevations between the first and second depth positions, and interference boundaries are no longer presented in explanatory text but are instead embedded in the templates themselves.

[0041] Second, templates are broken down based on location differences. The aforementioned rack templates are further linked to end aisles, middle aisles, side racks, back-to-back racks, and semi-double-depth scenarios. For end aisle templates, the stacker crane end buffer zone boundary and end clearance zone boundary are additionally recorded; for middle aisle templates, the aisle clear width, back-to-back distance, and double-sided service boundaries are recorded; for side rack templates, the back-pull space boundary between the outer side of the rack and the factory boundary is recorded, and this is used to constrain the installation safety distance of side racks in scenarios where they are arranged against a wall or edge; for back-to-back rack templates, the back-to-back spacing boundary between adjacent rack rows is recorded, and this is used to constrain the installation spacing and operational safety boundary in scenarios where two rows of racks are arranged in opposite directions; for semi-double-depth templates, the combination relationship between the single-depth service boundary on one side and the double-depth service boundary on the other side is recorded.

[0042] Third, install rule tags are attached to the templates. Each template is associated with an install rule tag, clearance boundary parameters, installation hole boundary parameters, and equipment operating boundary parameters, and its applicable depth, layer, location category, serviceable equipment category, and reverse correction priority are recorded. In other words, this step outputs not a single dimension, but a template object containing "geometric dimensions + rule tags + operating boundaries + subsequent correction interfaces." After this processing, the subsequent first layer is no longer optimized directly from an unconstrained parameter space, but rather engineering installability propagation and screening are performed at the template level.

[0043] Therefore, the purpose of this step is to transform the original engineering parameters into a calculable, determinable, screenable, and retrospectively correctable structural template, thereby truly implementing "hard constraint pre-positioning" and "differentiated installation rules" at the algorithm entry point.

[0044] S3. First-layer project installability generation mechanism based on constraint propagation and pruning rules: The function of this step is to: aim at "generating single-row shelving templates", rather than simply finding some feasible solutions, but to construct an "engineering installability generation mechanism with constraint propagation and pruning rules", which progressively shrinks the number of columns, layers, template type, length boundary and height boundary, and outputs only a set of structural templates that meet the rules of installation, safety and equipment operation.

[0045] To better illustrate the first-level engineering installability generation mechanism, the target planning model and engineering installability template set expression for single-depth and double-depth racks will be further explained below.

[0046] like Figure 7 As shown, in this embodiment, the first layer does not directly traverse all combinations of column and layer numbers. The first layer optimization establishes target planning models for single-deep and double-deep shelves respectively, with the number of shelf columns and shelf layers as decision variables, and all decision variables are limited to positive integers.

[0047] 1. Optimization model for the first level of a single-depth shelving unit: The goal is to minimize space wastage along the length and height of the factory building, expressed as: ; The constraints are: ; ; ; ; , Represents the set of positive integers; In the formula, This represents the objective function value for optimizing the first layer of a single-depth shelving unit. and These represent the total length and total height occupied by a single-depth rack or a single row of racks, respectively. and These represent the length and height of the factory building that can be used for stacker crane automated warehouse layout; This indicates the length of space occupied by one side of the stacker crane's end buffer. Indicates the height of the safety space at the bottom of the shelf; Indicates the width of the storage compartment; Indicates the height of the top shelf compartment in a single-depth shelving unit; Indicates the height of non-top-level shelves in a single-depth shelving unit; , These represent the number of rows and layers of a single-depth shelving unit, respectively, and both are positive integers.

[0048] 2. Optimization model for the first level of double-depth shelving: The goal is to minimize space wastage along the length and height of the factory building, expressed as: ; The constraints are: ; ; ; ; ; In the formula, This represents the objective function value for optimizing the first layer of a double-deep shelving unit. and These represent the total length and total height occupied by a single row of double-depth shelving units, respectively. This indicates the height of the top shelf compartment in the deepest position of a double-deep shelving unit. This indicates the height of the first, non-top-level shelf compartment in a double-deep shelving unit. , These represent the number of rows and layers of double-deep shelving, respectively, and both are positive integers.

[0049] 3. Project template set representation: After solving the first layer of optimization, the single-deep rack template and double-deep rack template that satisfy the constraints are represented as follows: ; ; Furthermore, the set of installable templates for the first-level output project can be represented as: ; In the formula, and These represent the sets of templates that can be installed in single-depth racking projects and double-depth racking projects, respectively. , This represents the optimal number of columns and optimal number of layers for a single-depth shelving unit obtained from the first level of optimization. , This represents the optimal number of columns and optimal number of layers for the double-deep shelving obtained from the first level of optimization; , This represents the total occupied length and total occupied height corresponding to the optimal single-depth rack solution; , This indicates the total occupied length and total occupied height corresponding to the optimal solution for double-depth shelving; , Indicates the installation rule label; , Indicates a template index; , This indicates the reverse correction interface.

[0050] In this embodiment, the first layer does not directly traverse all column and layer combinations, but performs constraint propagation and boundary shrinkage according to the following hierarchy: (1) Propagation along the length boundary: For single-depth templates, according to the formula Available length of the factory building Stacker crane end buffer space and compartment width Nexus Propagation yields an upper bound on the number of available columns for a single-depth template. .like If the result is negative, it indicates that the single-column, single-layer scheme can no longer be installed, and the corresponding single-depth template should be completely removed.

[0051] For double-depth templates, according to the formula Similarly, the upper bound of the number of available columns for a double-deep template can be obtained. If the combined width of the end buffer and the double-deep template exceeds the length boundary, the corresponding double-deep template will be completely removed.

[0052] (2) Propagation along the height boundary: For single-depth templates, according to the formula Useful height of the factory building Bottom safety space Top-level template height Non-top template height number of layers Propagation yields an upper bound on the number of available layers for a single-depth template. .when When the top-level template exceeds the boundary, the corresponding single-depth template is completely cut off.

[0053] For double-depth templates, according to the formula The factory building height boundary, bottom safety space, and the height of the top floor and non-top floor in the first deep section of the double-deep building are considered in terms of the number of floors. Propagation yields an upper bound on the number of available layers for the double-deep template. If the double-deep template cannot meet the height boundary in either the top layer or non-top layer direction, then the corresponding template should be cut.

[0054] (3) Propagation of hole location and template type: For all non-top-floor templates, the height of the storage compartments must be aligned with the center-to-center distance of the column holes. If a template, after rounding up, meets the geometric height requirements but would cause incompatibility with the installation elevation of the beams on adjacent floors, the hole sequence, or the top-floor clearance boundary, then that template will not be included in the candidate set. In other words, the column hole distance is not only involved in the calculation of the height of a single storage compartment but also in the determination of the propagation of installation continuity between floors.

[0055] (4) Deep-seated differential propagation: For double-deep templates, it is determined not only whether the first and second deep positions meet the height requirements for the top and non-top floors respectively, but also whether they match the operating depth of the telescopic fork of the stacker crane to be served. When the second deep template still cannot meet the operational safety requirements after deducting the extra lifting space of the double-deep template in the top-floor scenario, the second deep top-floor template is removed separately; when the combination of the first and second deep templates would create a service conflict within the same rack row, the corresponding double-deep template combination is removed as a whole.

[0056] (5) Location-difference propagation: For end aisle templates, additional requirements for end buffer space, end clearance, and track end boundaries must be met. For middle aisle templates, double-sided service boundaries, back-to-back spacing, and back-pull space must be met. For side rack templates, back-pull space constraints on the outer side of the racks must be met. For back-to-back rack templates, back-to-back spacing constraints between adjacent rack rows must be met. For semi-double-depth templates, the combined service rules of single-depth on one side and double-depth on the other side must be met simultaneously. If the boundary conditions of the template at the corresponding location are not met, it will not be included in the candidate set.

[0057] (6) Pruning output: After the above propagation process, only templates that meet the rules for factory boundaries, hole positions, depth, floor levels, end points, back tension, fire protection, and service boundaries are retained. Each template is then labeled with an installation rule tag, a template index, and a reverse correction interface, forming the first-level output set of installable templates for the project. Here It is no longer a set of feasible solutions in the general sense, but rather an "engineering template pool after step-by-step rule propagation and pruning".

[0058] Therefore, the purpose of this step is to shrink the search objects of the subsequent second layer from the original parameter space to the project installable template pool, reduce large-scale invalid searches, and at the same time move the installation logic forward, highlighting the "candidate solution generation and screening mechanism oriented towards project installation rules".

[0059] S4. A second-layer dedicated solution mechanism for repairing projections of discrete values ​​in the device catalog and illegal solutions: The function of this step is to install the template set in the project output at the first level. Furthermore, multi-objective collaborative optimization is carried out on the number of roadways, roadway combinations, and stacker crane speed parameter combinations. In the optimization process, the discrete value constraints of the equipment catalog, the illegal solution repair projection of the current layer, and the identification / reverse correction of local installation conflicts are truly written into engineering mechanisms, rather than just as penalty terms of general genetic algorithms.

[0060] like Figure 8 As shown, in this embodiment, the second-layer dedicated solution mechanism for S4, which addresses the discrete values ​​of the equipment catalog and the illegal solution repair projection, includes the following steps: First, input the set of installable templates for the project, equipment catalog parameters, project parameters, and boundary conditions output from the first layer; second, initialize the population based on the set of installable templates for the project, where each individual includes a template index gene, aisle quantity gene, equipment index gene, and position constraint flags and correction flags; then, calculate the objective function and check the constraints, where the objective function includes the space waste objective function in the plant width direction, the comprehensive cost objective function, and the total operating efficiency objective function of all aisle stacker cranes, and the constraints include the number of storage locations constraint, operating efficiency constraint, aisle width constraint, plant width boundary constraint, and semi-double-deep stacker crane efficiency constraint; when the equipment index in an individual does not meet the equipment catalog constraints or installation boundary constraints, perform equipment catalog constraint projection repair. The solution is mapped to the nearest valid combination of equipment. Then, local installation conflict identification is performed to determine if candidate solutions have issues such as end interference, insufficient clearance, misaligned mounting holes, mismatched top-level and non-top-level rules, mismatched first-depth and second-depth rules, inconsistent equipment operating boundaries, or mismatched position templates. If a conflict exists and can be repaired at the current layer, repair is performed at the current layer. If the conflict cannot be repaired at the current layer, reverse correction is performed and the process backtracks to the first-layer set of installable templates. After repair or reverse correction, selection, crossover, and mutation are performed, and the next generation population is generated through elite retention. Finally, termination conditions are checked. Termination conditions include reaching the preset maximum number of iterations or no significant improvement in the Pareto front within a preset number of generations. If the termination conditions are met, the Pareto optimal solution set is output. If the termination conditions are not met, the process returns to the steps of calculating the objective function and checking constraints to continue iteration.

[0061] To facilitate the explanation of the second-layer dedicated solution mechanism, the decision variables, objective function, constraints, and equipment catalog projection repair machine manufacturing are further explained below.

[0062] (1) Multi-objective optimization objective: In this embodiment, the decision variables for the second-level optimization include integer decision variables and discrete decision variables. The discrete decision variables include the combination of single-depth stacker crane speed parameters. Combined speed parameters with double-deep stacker crane ,in: ; ; Furthermore, ; ; In the formula, These represent the number of aisles for single-deep stacker cranes, double-deep stacker cranes, and semi-double-deep stacker cranes, respectively. Represents the set of non-negative integers; This indicates the combination of speed parameters for a single-depth stacker crane. This indicates the combination of speed parameters for a double-deep stacker crane; This represents a preset discrete set of values ​​for the speed parameter combinations of a single-depth stacker crane. This represents a preset discrete set of values ​​for the speed parameter combinations of a double-deep stacker crane; This indicates the travel speed of a single-depth stacker crane; This indicates the lifting speed of a single-depth stacker crane; This indicates the speed of the telescopic fork of a single-depth stacker crane; This indicates the travel speed of the double-deep stacker; This indicates the lifting speed of the double-deep stacker; This indicates the speed of the telescopic fork of a double-deep stacker.

[0063] 1. Objective function: Minimizing space waste in the width direction of the factory building is expressed as: ; The minimum overall cost is expressed as: ; The maximum overall operating efficiency of all stacker cranes in all aisles is expressed as: ; In the formula, This represents the objective function value indicating the waste of space in the width direction of the factory building; This represents the value of the overall cost objective function; This represents the objective function value for the overall operating efficiency of stacker cranes in all aisles; This indicates the width of the factory building that can be used for stacker crane automated warehouse layout; Indicates the width of a single-depth tunnel; Indicates the width of a double-deep tunnel; Indicates the width of a semi-double-deep tunnel; This indicates the back-to-back spacing between shelves in adjacent aisles; Indicates the space required for back-pull shelving; This indicates the speed parameter combination of a single-depth stacker crane. The corresponding equipment cost; This indicates the speed parameter combination of the double-deep stacker crane. The corresponding equipment cost; This indicates the cost per storage location; This indicates the total number of storage locations in the plan; This indicates the total length occupied by a single-depth rack or a single row of racks; This indicates the total height occupied by a single-depth rack or a single row of racks; This indicates the total length occupied by a single row of double-deep shelving; This indicates the total height occupied by a single row of double-deep shelving; This indicates the speed parameter combination of a single-depth stacker crane. Total length occupied by a single row of shelves Total height occupied by single-row shelving Operational efficiency under certain conditions; This indicates the speed parameter combination of the double-deep stacker crane. Total length occupied by a single row of shelves Total height occupied by single-row shelving Operational efficiency under certain conditions; This indicates the operational efficiency corresponding to a semi-double-depth tunnel.

[0064] 2. Constraints: The quantity constraint for storage locations is: ; ; The work efficiency constraint is: ; The width constraint of the tunnel is: ; ; ; The factory building width boundary constraint is: ; The efficiency constraint for a semi-double-deep stacker is: ; In the formula, Indicates the efficiency of project requirements; Indicates the number of storage locations in a single storage compartment; This represents the optimal number of columns for a single-depth shelving unit obtained from the first level of optimization. This represents the optimal number of single-depth shelves obtained from the first level of optimization; This represents the optimal number of columns for the double-deep shelving obtained from the first level of optimization; This represents the optimal number of layers for the double-deep shelving obtained from the first layer of optimization; Indicates the minimum number of storage locations required for the project; This indicates the total number of cargo units corresponding to a single-depth aisle; This indicates the total number of cargo units corresponding to a double-deep aisle; This indicates the total number of cargo units corresponding to a semi-double-depth aisle; This indicates the distance between materials on the loading platform and materials on the shelves in the aisle; Indicates the length of a single stored material; This indicates the distance between the first and second deep items in a double-deep shelving unit.

[0065] 3. Equipment catalog projection repair expression: To embody the "dedicated solution mechanism for repairing discrete values ​​in the equipment catalog and illegal solutions in engineering," this embodiment can further represent the projection repair of illegal equipment combinations as follows: ; in, This represents the set of legal equipment combinations under the current template, lane width, service height, and service length constraints. Indicates illegal device combination Combined with legal equipment The comprehensive distance function between them; the comprehensive distance function includes at least the minimum net width difference of the roadway, the service height difference, the service length difference, and the cost increment.

[0066] When the device index in an individual does not meet the device catalog constraints or installation boundary constraints, the projection operator is used. Instead of directly declaring it invalid, it is mapped to the most recent legitimate combination of devices.

[0067] In this embodiment, the second layer includes at least the following objectives: first, to minimize wasted space in the width direction of the plant; second, to minimize the combined cost of stacker cranes and storage locations; and third, to maximize the overall operating efficiency of stacker cranes in all aisles. This simultaneously satisfies constraints on the number of storage locations, efficiency, aisle width, plant width boundary, and semi-double-depth efficiency.

[0068] The individual encoding in this embodiment is not a simple "integer encoding + discrete encoding", but consists of the following parts: the first part is a template index gene, used to extract from... The first part selects single-depth and double-depth templates; the second part is the roadway quantity gene. The first part uses integer encoding; the second part is the device index gene. ,in Points to a specific actual model in the single-depth equipment catalog. The first part points to a specific real model in the dual-deep equipment catalog; the fourth part consists of position constraint flags and correction flags, used to record the current individual's end occupancy, intermediate roadway type, back-to-back combination type, and whether it has been projected and repaired. Through this encoding method, each individual in this embodiment is bound to the real equipment catalog and installable templates from the beginning of its generation, thereby avoiding the generation of a large number of "virtual equipment combinations" that do not exist in reality using general continuous variables.

[0069] After crossover, mutation, or template replacement, if an individual exhibits any of the following conditions: the selected speed parameter combination is not within the legal value range of the equipment catalog; the selected equipment service height is insufficient to cover the template height; the selected equipment service length is insufficient to cover the length occupied by a single row of shelves; the minimum aisle net width required by the selected equipment is greater than the aisle net width of the current scheme; or the selected double-deep equipment does not support the current double-deep first and second depth template combinations, then a penalty value is not simply applied; instead, equipment catalog constraint projection repair is triggered. Specifically, the system repairs in the following order: first, it determines whether there is a nearest legal equipment when the template type remains unchanged; if so, it projects the illegal equipment index to the nearest installable equipment combination based on the comprehensive distance of "minimum net width difference + minimum height difference + minimum length difference + minimum cost increment"; if not, it keeps the equipment index unchanged, reverses the template index, and selects the nearest template compatible with the current equipment; if neither the template nor the equipment can be projected into a legal combination through the current layer, it proceeds to the reverse correction process. The result of this process is that illegal individuals are prioritized for repair to the nearest installable individual in the project, rather than being directly discarded.

[0070] For each generation of individuals, local installation conflict identification is performed before calculating the objective function. Conflicts include at least the following types: end interference, insufficient clearance, misaligned mounting holes, mismatch between top-level and non-top-level rules, mismatch between first-depth and second-depth rules, inconsistent equipment operating boundaries, and mismatched position templates. In this embodiment, the identification method is not a simple Boolean judgment, but rather the formation of conflict vectors, such as length exceeding limits, height exceeding limits, insufficient roadway width, mounting hole deviation, equipment catalog compatibility flag, deep-level service conflict flag, and end-buffer boundary conflict flag. Minor conflicts are processed at the current layer, while moderate to severe conflicts trigger backtracking.

[0071] When the conflict is minor, such as the number of lanes exceeding one position, equipment index deviating from the nearest installable combination, or location template mismatch that can be repaired by replacing with a template of the same category, this embodiment prioritizes repairing the current layer. The repair order can be set as follows: repair the equipment index first, then the template index, and finally the number of lanes. For example, when the total width occupied by the lanes slightly exceeds the width of the factory building, a reduction should be attempted first. Alternatively, some semi-double-deep lanes may be replaced with single-deep lanes; when the equipment operating boundaries do not match, the equipment index shall be projected to the nearest valid model in the same directory; when the location template is mismatched, the end template shall be replaced with the middle template or vice versa, but replacement is only allowed when the rule base constraints are met.

[0072] When a conflict is classified as moderate to severe and cannot be resolved through the current layer repair mechanism—for example, a conflict between the template's height combination and the factory boundary, a complete mismatch between the double-deep top-level template and the second-deep service boundary, the absence of a compatible and valid model in the equipment catalog, or an inability to meet the factory width and efficiency minimum requirements regardless of adjustments to the number of lanes under the current template—a reverse correction mechanism is triggered. In this embodiment, the reverse correction rule module executes at least the following backtracking logic: prioritizing backtracking the template layer number. or Secondly, backtrack the number of template columns. or The process involves backtracking the top-level template type, deep-level template type, or location-based template type; if necessary, backtracking and regenerating the corresponding template category. After backtracking, the corrected template is sent back to the second layer for multi-objective evaluation, instead of starting the search globally again. This preserves previous optimization information while avoiding the inefficient handling of "the entire solution being invalidated after a local conflict" in traditional solutions.

[0073] Therefore, the purpose of this step is to truly upgrade the second-level optimization from a "general improved genetic algorithm" to a "dedicated solution mechanism for repairing illegal solutions in real-world equipment catalogs, installation boundaries, and engineering processes".

[0074] S5. Generation of installation-level parameter package: The function of this step is to select the target solution from the Pareto optimal solution set and transform the target solution into an installation-level parameter package that can directly support construction, installation, and delivery. This step is not simply exporting the optimization results, but rather converting the optimization solution into a set of engineering implementation parameters.

[0075] In this embodiment, the installation-level parameter package includes at least the following: 1. The geometric parameters of the storage compartments corresponding to single-depth / double-depth / semi-double-depth templates; 2. Number of columns, number of layers, top-level template type, and non-top-level template type for each row of shelves; 3. The sequence of hole positions, beam installation elevation, and shelf height for each shelf upright; 4. Centerline coordinates, net width, service boundary, and track centerline location for each tunnel; 5. Equipment model, speed parameters, service height, service length, and compatible aisle number for each stacker crane; 6. Stacker crane end buffer zone boundary, fire clearance boundary, back-pull boundary, and back-to-back boundary; 7. Interference verification results of the first / second depth positions of the double-deep telescopic fork; 8. Corresponding equipment BOM, shelving BOM, and installation list.

[0076] It is particularly important to emphasize here that the installation hole sequence, beam installation elevation, track centerline position, stacker crane buffer zone boundary, fire clearance zone boundary, double-deep telescopic fork interference verification results, and equipment BOM / installation list are all automatically generated by the aforementioned template, equipment catalog, constraint propagation results, and second-level optimization results, rather than being manually entered. Thus, the output is no longer a "layout display diagram" but an "installation-level parameter package."

[0077] S6. Installation-level parameter package consistency verification and project closed-loop output: The function of this step is to perform a consistency check on the installation-level parameter package generated in step S5 before project delivery, and after the check passes, drive the drawing generation system to output the storage unit diagram, floor plan, elevation layout, stacker crane parameter table, installation hole table, clearance check table and equipment bill of materials table, thereby forming a closed-loop output chain of "installable scheme - installation-level parameter package - construction delivery drawings".

[0078] In this embodiment, the consistency check is performed in at least the following order: First, layout-equipment consistency verification. Verify whether the stacker crane model and speed parameters match its service aisle width, service height, and service length; Second, verify the consistency between floor height and hole position. Verify whether the installation elevation of the beams on each floor strictly corresponds to the sequence of column hole positions; Third, verify the consistency of clearance and safety boundaries. Verify whether the top clearance, bottom clearance, fire clearance, end buffer zone boundaries, and back tie space meet the requirements of the rule base; Fourth, double-depth interference consistency verification. Verify whether there is interference between the first / second depth template of the double-depth system and the working boundary of the telescopic fork of the selected equipment; Fifth, verify the consistency between drawings and the Bill of Materials (BOM). Verify that the quantities of shelves, equipment, rail lengths, beams, and installation components in the drawings match those in the BOM. Sixth, rule label-delivery result consistency verification. Verify whether the final output drawings still correspond to the original installation rule labels and the template categories after retrospective correction.

[0079] Only after all verifications pass will the system send the installation-level parameter package to the drawing generation system, which will then automatically output the following engineering deliverables: storage unit diagram, automated warehouse floor plan, automated warehouse elevation layout, stacker crane parameter table, installation hole location table, clearance check table, equipment and racking BOM, and installation list.

[0080] Among them, the key dimensions, hole positions, track center lines, buffer zone boundaries, fire clearance zone boundaries, and double-deep interference verification results in the drawings all originate from the final solution after the aforementioned constraint propagation, template generation, equipment catalog projection repair, and reverse correction, forming a true engineering closed-loop output. In other words, the constraint-coupled stacker crane automated warehouse layout generation method in this project is not "drawing with optimization results," but rather "directly generating construction and installation parameter packages from the engineering installable scheme and outputting delivery drawings."

[0081] Furthermore, to verify the feasibility, effectiveness, and engineering applicability of the stacker crane automated warehouse layout generation method with engineering constraint coupling described in this invention, this embodiment selects a set of actual engineering parameters for testing. The testing process includes engineering parameter input, parametric modeling of storage cell units, generation of the first-layer installable template, second-layer multi-objective collaborative optimization, generation of installation-level parameter packages, and drawing output verification.

[0082] I. Input parameters: In this embodiment, the input parameters are as follows: The usable length of the factory building is 96,000 mm; The usable width of the factory building is 18500mm; The usable height of the factory building is 16000mm; The material unit length is 1200mm; The width of the material unit is 1000mm; The height of the material unit is 1600mm; Number of storage compartments per storage location It is 2; The width of the shelf uprights is 100mm; The center-to-center distance of the upright holes in the shelving unit is 75mm. The crossbeam height is 100mm; The distance between materials is 100mm; The material is 75mm away from the column; The material should be 100mm above the safety distance of the crossbeam. The safe lifting space is 150mm; The fire protection clearance is 150mm. The safety clearance at the bottom of the shelf is 800mm; The top safety space for a single-depth top layer is 800mm; The top safety space of the double-deep top layer is 1000mm; The additional lifting space for the double depth is 150mm; The distance between the loading platform and the materials on the shelves in the aisle is 150mm. The distance between the first and second deepest materials in the double-deep system is 50mm. The back-to-back spacing is 400mm; The required space for the back strap is 350mm. The end buffer space of the stacker crane is 3500mm; The candidate speed combinations for a single-deep stacker crane are: travel speed 140 / 160m / min, lifting speed 30 / 45m / min, and telescopic fork speed 30 / 45m / min. The candidate speed combinations for the double-deep stacker crane are: travel speed 140 / 160m / min, lifting speed 30 / 45m / min, and telescopic fork speed 30 / 45m / min. The minimum number of storage locations required is no less than 4,000. The system's operating efficiency requirement is no less than 100 torts / hour; The price for a single storage space is 380 yuan.

[0083] The prices of stacker crane equipment are as follows: single-deep stacker cranes are priced from RMB 450,000 to RMB 540,000 under different speed combinations, and double-deep stacker cranes are priced from RMB 500,000 to RMB 590,000 under different speed combinations. The specific values ​​are based on discrete values ​​from the preset equipment catalog.

[0084] II. Implementation Process: (a) Parametric modeling of storage cell elements with coupled engineering constraints: Based on the input engineering parameters, parametric models of single-depth and double-depth shelving units are first constructed.

[0085] 1. Calculation of unit dimensions for single-depth shelving units: The width of a single-depth shelving unit is calculated using the following formula: ; Substituting the parameters into this embodiment, we get: ; The length of a single-depth shelving unit is calculated using the following formula: ; Substituting the parameters into this embodiment, we get: ; The height of non-top-level shelves in a single-depth rack is calculated using the following formula: ; Substituting the parameters into this embodiment, we get: ; The height of the top shelf compartment of a single-depth shelving unit is calculated using the following formula: ; Substituting the parameters into this embodiment, we get: ; 2. Calculation of unit dimensions for double-deep shelving: Double-deep shelving units have the same width and length as single-deep shelving units, therefore: , ; The height of the non-top shelf compartment in the first and second deep positions of double-deep shelving is calculated using the following formula: ; Substituting the parameters into this embodiment, we get: ; The height of the top shelf compartment in the deepest position of a double-deep shelving unit is: ; The height of the top shelf compartment in the second deep position of a double-deep shelving unit is: ; After obtaining the above basic dimensions, this embodiment further generates a single-depth non-top-level template, a single-depth top-level template, a double-depth first-depth non-top-level template, a double-depth second-depth non-top-level template, a double-depth first-depth top-level template, and a double-depth second-depth top-level template, and respectively attaches a layer label, depth label, clearance boundary label, installation hole boundary label, and equipment operation boundary label, as input templates for the subsequent first-layer engineering installability generation mechanism.

[0086] (II) First-level optimization: Template generation for engineering installation based on constraint propagation and pruning rules: In this embodiment, the first layer of optimization does not simply determine the number of columns and layers, but rather, based on length boundaries, height boundaries, column hole spacing, layer differences, depth differences, end rules, and fire clearance rules, it constrains and prunes the templates, retaining only the installable templates.

[0087] 1. Single-depth racking and single-row racking target planning: The optimization model for the first layer of a single-depth shelving unit is as follows: ; The constraints are as follows: ; ; ; ; ; Propagation along the length direction reveals that the maximum number of columns in a single-depth template satisfies: ; therefore, .

[0088] Propagation along the height direction reveals that the maximum number of layers in a single-depth template satisfies: therefore, .

[0089] By combining the goal programming solution, the optimal template parameters for a single-depth shelving unit are obtained as follows: number of columns. number of floors The total length occupied by a single row of shelves is: ; The total height occupied by a single row of shelves is: ; It is evident that the above results simultaneously meet the factory building length and height boundaries, and the requirements for top-level formwork, non-top-level formwork, hole pitch, and clearance all meet the installation conditions.

[0090] 2. Target planning for double-depth single-row shelving: The optimization model for the first layer of a double-deep shelving unit is as follows: ; The constraints are as follows: ; ; ; ; ; Similarly, propagation along the length direction yields the following result. Obtained through propagation in the vertical direction: ; therefore .

[0091] By combining the goal programming solution, the optimal template parameters for the double-deep shelving are obtained as follows: number of columns. number of floors The total length occupied by a single row of shelves is: ; The total height occupied by a single row of shelves is: ; The results also meet the requirements for the length, height, and top, non-top, and deep-position operation boundaries of the factory building and double-deep formwork.

[0092] 3. First layer output: After the first layer of processing, the output of this embodiment is not a feasible solution in the ordinary mathematical sense, but a template set containing only installable templates for the project. Each template in this set comes with an installation rule label, a template index, and a reverse correction interface, serving as the sole input basis for the second layer of optimization. This avoids a large number of ininstallable candidate solutions from the full-range search entering the second layer of optimization.

[0093] (III) Second-level optimization: Multi-objective collaborative optimization and equipment catalog discrete value projection repair: In this embodiment, the set of installable templates generated in the first layer is used as the sole basis for solving a multi-objective optimization model, which is then solved using a dedicated improved genetic algorithm within the NSGA-II framework. Integer decision variables include the number of single-depth tunnels. Double-deep alleyway and half-double deep alleyway Discrete decision variables include combinations of single-depth stacker crane speed parameters. Combined speed parameters with double-deep stacker crane And only discrete selection from the actual device catalog is allowed.

[0094] 1. Objective function and constraints: The second-level optimization objectives include: First, minimize the waste of space in the width direction of the factory building; Second, minimize overall costs; Third, maximize the overall operating efficiency of stacker cranes in all aisles.

[0095] The quantity constraint for storage locations is: ; Substituting into this embodiment , , , , Then, the program calculates the total number of storage locations under different combinations of lane numbers and filters out candidate solutions with fewer than 4,000 storage locations.

[0096] The efficiency constraint is: ; The efficiency of single-depth and double-depth operations is calculated according to the preset efficiency model, while the efficiency of semi-double-depth operations is calculated according to the combined conversion rules.

[0097] 2. Equipment catalog discrete value projection repair: If the following situations occur during crossover, mutation, and individual update: The selected speed combination is not within the range of valid discrete values ​​in the equipment catalog; The selected equipment is not sufficiently serviced. The selected equipment has insufficient service length; The minimum clear width of the roadway required by the selected equipment exceeds the clear width of the current plan; The selected double-depth equipment is incompatible with the double-depth template; The system will not directly classify the individual as invalid. Instead, it will automatically map the illegal device index to the most recently installable device combination based on the device catalog constraint projection repair mechanism.

[0098] In this embodiment, the device speed parameters corresponding to the optimized output target solution are: The speed combination for a single-depth stacker crane is [160, 45, 45]; The speed combination of the double-deep stacker is [160, 45, 45].

[0099] The results show that the program ultimately selected a high-speed combination of single-deep and double-deep stacker cranes from the actual equipment catalog, and through discrete catalog constraint projection and boundary repair, made it consistent with the template height, template length, aisle net width and efficiency targets.

[0100] 3. NSGA-II Optimization Results In this embodiment, the target solution obtained by the second layer of optimization is: Number of single-depth tunnels ; Double-deep tunnel number ; Half-double deep alleyway number .

[0101] The corresponding total number of storage locations is: ; It is evident that this solution meets the requirement of a minimum of 4,000 storage locations.

[0102] According to the program output, the total width occupied by this solution is 17100mm, with a remaining width space of 1400mm; the total cost is 3.2947 million yuan; and the total efficiency is 116.64 pallets / hour. Therefore, this solution simultaneously satisfies the factory width boundary constraints, the number of storage locations constraints, and the system efficiency constraints. The optimization results you provided also show that: the speed parameters are "single depth [160, 45, 45], double depth [160, 45, 45]", the feasible solution is "number of single-depth aisles = 1, number of double-depth aisles = 1, number of semi-double-depth aisles = 1", the total width occupied is 17100mm, the total number of storage locations is 4144, the remaining width space is 1400mm, the total cost is 3.2947 million yuan, and the total efficiency is 116.64 pallets / hour.

[0103] 4. Significance of the results: The test results show that this invention does not simply pursue the maximum number of cargo bays or the lowest cost, but rather achieves a trade-off optimization between space, cost, and efficiency while satisfying engineering constraints, equipment catalog constraints, and installation boundary constraints. Specifically, if candidate solutions encounter problems such as equipment incompatibility, insufficient aisle width, or local installation conflicts in the second layer, these issues will be addressed first through current layer repair and equipment catalog projection repair. If these cannot be repaired, a reverse correction is triggered, backtracking the template parameters before solving again, thereby improving the engineering feasibility of the Pareto optimal solution set.

[0104] (iv) Generation and verification of installation-level parameter packages and drawing output: After obtaining the above target solution, this embodiment further extracts layout parameters, equipment selection parameters, mounting hole parameters, clearance parameters, safety boundary parameters, and bill of materials parameters to form an installation-level parameter package.

[0105] This installation-level parameter package includes at least: Cargo compartment size parameters corresponding to single-depth, double-depth, and semi-double-depth templates; Shelf column number, number of layers, and top and non-top shelf template information; Shelf upright hole sequence and beam installation elevation; Location of the tunnel centerline, tunnel clearance width, and track centerline; Single-deep and double-deep stacker crane models, speed parameters, and service boundaries; Stacker crane end buffer zone boundary; Fire safety clearance boundary, top clearance boundary, and bottom clearance boundary; Interference verification results of the first and second depth positions of the double-deep telescopic fork; Equipment BOM, shelving BOM, and installation list.

[0106] In this embodiment, the system inputs the aforementioned installation-level parameter package into the parametric drawing generation system to automatically generate a CAD layout drawing of the automated warehouse, such as... Figure 9 As shown.

[0107] III. Test Result Analysis: The actual test results of this embodiment show that: First, the present invention can reliably generate a grid template and layout scheme that meet the engineering installation conditions after inputting the actual factory building size, rack size, stacker crane discrete equipment catalog and project requirement constraints; Second, the present invention eliminates templates that do not meet the length, height, hole position, depth and position boundaries in advance through the first layer of constraint propagation and pruning, thereby reducing the invalid search range of the second layer of optimization; Third, the present invention uses the second-layer equipment catalog discrete value projection repair and local conflict repair mechanism to couple the optimization process with the actual equipment selection and actual installation boundary, thus avoiding a large number of candidate solutions that are only mathematically feasible but not engineeringally feasible from entering the final result; Fourth, the target solution output by this invention meets the design requirements of having no less than 4,000 storage locations, a system efficiency of no less than 100 pallets / hour, and a factory width that does not exceed the limit. Furthermore, it can generate installation-level parameter packages and engineering drawings, demonstrating that this invention has good engineering practicality and feasibility.

[0108] In summary, this embodiment verifies the effectiveness of the method of the present invention through a set of actual engineering parameters. The test results show that the present invention can achieve automatic generation, collaborative optimization solution, and installation-level closed-loop output of the layout scheme of the stacker crane automated storage and retrieval system, under the premise of meeting the constraints of plant boundary, rack installation boundary, equipment catalog, efficiency constraints, and delivery output requirements. This proves that the present invention has clear engineering application value and feasibility.

[0109] The above are merely preferred embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A method for generating the layout of a stacker crane automated warehouse with engineering constraint coupling, characterized in that, Includes the following steps: S1. Collect the available space parameters of the factory, material unit parameters, rack parameters, stacker crane parameters and project requirement parameters required for the design of the stacker crane automated warehouse, and establish an installability rule library based on the collected parameters; S2. Based on the parameters and installability rule base in step S1, perform parametric modeling of single-deep racking unit and double-deep racking unit, and generate corresponding installable templates according to different depths, different levels and different aisle locations. S3. With the goal of minimizing space waste in the length and height directions of the factory building, and with constraints such as the stacker crane end buffer space, the safety space at the top and bottom of the rack, the matching of the upright hole spacing, the fire protection reserved space, and the factory boundary, the first-level optimization is carried out on single-deep racks and double-deep racks respectively. The number of columns and layers of a single row of racks is solved, and a set of feasible templates with installation rule labels and reverse correction interfaces is generated. S4. Using only the feasible template set generated in step S3 as the solution basis, and with multiple objectives of minimizing space waste in the width direction of the factory, minimizing overall cost, and maximizing operational efficiency, a second layer of optimization is performed on the combination of single-depth aisle number, double-depth aisle number, semi-double-depth aisle number, and stacker crane speed parameters. During the iterative optimization process, local installation conflict identification is performed on the generated candidate solutions. When there is a repairable conflict, illegal solution repair is performed according to the installability rule base. When there is an installation conflict that cannot be directly repaired in the current layer, a reverse correction mechanism is triggered, backtracking and correcting the feasible template parameters in step S3, and then re-evaluating the optimization to finally obtain the Pareto optimal solution set. S5. Select target solutions from the Pareto optimal solution set in step S4, and extract layout parameters, equipment selection parameters, installation hole parameters, clearance parameters, safety boundary parameters and bill of materials parameters to form an installation-level parameter package; S6. Perform a consistency check on the installation-level parameter package in step S5. After the check passes, drive the drawing generation system to output the layout drawings and parameter list of the automated warehouse that can be directly used for engineering installation. The installability rule base established in step S1 includes at least the following: geometric installation rule sub-base, layer-level differentiation rule sub-base, depth-level differentiation rule sub-base, location-level differentiation rule sub-base, reverse correction rule sub-base, and delivery consistency verification rule sub-base; The geometric installation rule sub-library constrains length, width, height, end buffer space, aisle clear width, back-pull space, and back-to-back spacing; the layer-level differentiation rule sub-library distinguishes the height and safety space requirements of top-level and non-top-level shelves; the depth-level differentiation rule sub-library distinguishes the size and operational boundary requirements of single-depth shelves, double-depth first-depth shelves, double-depth second-depth shelves, and semi-double-depth scenarios; the location-level differentiation rule sub-library distinguishes the installation templates for end aisles, middle aisles, side shelves, and back-to-back shelves; the reverse correction rule sub-library determines the backtracking object, correction order, and correction scope when installation conflicts are found in the second-level optimization; and the delivery consistency verification rule sub-library performs cross-checking of layout parameters, equipment selection parameters, installation hole parameters, clearance parameters, safety boundary parameters, and bill of materials parameters. The installable templates generated in step S2 include at least: single-depth non-top template, single-depth top template, double-depth first-depth non-top template, double-depth second-depth non-top template, double-depth first-depth top template, double-depth second-depth top template, end aisle template, middle aisle template, back-to-back shelf template, and semi-double-depth template; Each template is associated with the corresponding installation rule label, clearance boundary parameters, installation hole boundary parameters, and equipment operation boundary parameters.

2. The method for generating the layout of a stacker crane automated warehouse with engineering constraint coupling according to claim 1, characterized in that: In step S2, the parametric modeling of single-deep racking units and double-deep racking units includes at least: single-deep racking units and double-deep racking units; The width of a single-depth shelving unit is: ； The length of a single-depth shelving unit is: ； The height of non-top-level shelves in a single-depth rack is: ； The height of the top shelf compartment in a single-depth shelving unit is: ； The heights of the first and second deep non-top shelf compartments in double-deep shelving are as follows: ； The height of the top shelf compartment in the deepest position of a double-deep shelving unit is: ； The height of the top shelf compartment in the second deep position of a double-deep shelving unit is: ； The height of non-top-level shelves is calculated by rounding up and taking into account the center distance between the shelf upright holes. Alignment is performed to ensure that the height of the storage compartment matches the mounting holes on the shelf; In the formula, Indicates the width of the storage compartment; Indicates the length of the storage compartment; This indicates the height of the non-top shelf compartment in a single-depth shelving unit; Indicates the height of the top shelf compartment in a single-depth shelving unit; This indicates the height of the first, non-top-level shelf compartment in a double-deep shelving unit. This indicates the height of the second deep shelf unit (excluding the top shelf) in a double-deep shelving system. This indicates the height of the top shelf compartment in the deepest position of a double-deep shelving unit. This indicates the height of the top shelf compartment in the second deepest position of a double-deep shelving unit; , , These represent the width, length, and height of the stored material, respectively. Indicates the number of storage locations in a single storage compartment; Indicates the spacing between materials; Indicates the distance between the materials and the shelf uprights; Indicates the width of the shelf uprights; This indicates the safe distance that the material should extend beyond the crossbeam of the storage compartment; This indicates raising the safety space; Indicates space reserved for fire protection; Indicates the height of the beam; Indicates the center distance between the holes in the shelf uprights; Indicates the safety space at the top of a single-depth shelving unit; Indicates the safety space at the top of double-deep shelving; This indicates the additional lifting space required for double-deep shelving; This indicates the rounding up operation.

3. The method for generating the layout of a stacker crane automated warehouse with engineering constraint coupling according to claim 1, characterized in that: In step S3, the first layer of optimization establishes target programming models for single-deep and double-deep shelving respectively, with the number of shelving columns and the number of shelving layers as decision variables, and all decision variables are positive integers; The optimization model for the first level of a single-depth rack, aiming to minimize space waste in both the length and height directions of the factory building, is expressed as follows: ； The constraints are: ； ； ； ； , Represents the set of positive integers; The optimization model for the first level of double-deep shelving, aiming to minimize space waste in both the length and height directions of the factory building, is expressed as follows: ； The constraints are: ； ； ； ； ； After solving the first layer of optimization, the single-deep rack template and double-deep rack template that satisfy the constraints are represented as follows: ； ； in, This represents the objective function value for optimizing the first layer of a single-depth shelving unit. This represents the objective function value for optimizing the first layer of a double-deep shelving unit. and These represent the total length and total height occupied by a single-depth rack or a single row of racks, respectively. and These represent the total length and total height occupied by a single row of double-depth shelving units, respectively. and These represent the length and height of the factory building that can be used for stacker crane automated warehouse layout; This indicates the length of space occupied by one side of the stacker crane's end buffer. Indicates the height of the safety space at the bottom of the shelf; Indicates the width of the storage compartment; Indicates the height of the top shelf compartment in a single-depth shelving unit; This indicates the height of the non-top shelf compartment in a single-depth shelving unit; This indicates the height of the top shelf compartment in the deepest position of a double-deep shelving unit. This indicates the height of the first, non-top-level shelf compartment in a double-deep shelving unit. , , , These represent the number of rows, layers, and levels of a single-deep shelving unit, respectively, and are all positive integers. and These represent the sets of feasible templates for single-depth shelving and the sets of feasible templates for double-depth shelving, respectively. , This represents the optimal number of columns and optimal number of layers for a single-depth shelving unit obtained from the first level of optimization. , This represents the optimal number of columns and optimal number of layers for the double-deep shelving obtained from the first level of optimization; , This represents the total occupied length and total occupied height corresponding to the optimal single-depth rack solution; , This represents the total occupied length and total occupied height of the optimal double-depth shelving solution; , Indicates the installation rule label; , Indicates a template index; , This indicates the reverse correction interface.

4. The method for generating the layout of a stacker crane automated warehouse with engineering constraint coupling according to claim 1, characterized in that: In step S4, the decision variables for the second-level optimization include integer decision variables and discrete decision variables. The discrete decision variables include the combination of single-depth stacker crane speed parameters. Combined speed parameters with double-deep stacker crane ,in: ； ； ； ； The objective function of the second-level optimization should include at least the following: minimizing space waste in the width direction of the plant, minimizing the comprehensive cost consisting of stacker crane cost and storage location cost, and maximizing the total operating efficiency of stacker cranes in all aisles. Minimizing space waste in the width direction of the factory building is expressed as: ； The minimum overall cost is expressed as: ； The maximum overall operating efficiency of all stacker cranes in all aisles is expressed as: ； The constraints include: number of storage locations, operational efficiency, aisle width, plant width boundary, and semi-double-deep stacker crane efficiency. The quantity constraint for storage locations is: ； ； The work efficiency constraint is: ； The width constraint of the tunnel is: ； ； ； The factory building width boundary constraint is: ； The efficiency constraint for a semi-double-deep stacker is: ； In the formula, These represent the number of aisles for single-deep stacker cranes, double-deep stacker cranes, and semi-double-deep stacker cranes, respectively. Represents the set of non-negative integers; This indicates the combination of speed parameters for a single-depth stacker crane. This indicates the combination of speed parameters for a double-deep stacker crane; This represents a preset discrete set of values ​​for the speed parameter combinations of a single-depth stacker crane. This represents a preset discrete set of values ​​for the speed parameter combinations of a double-deep stacker crane; This indicates the travel speed of a single-depth stacker crane; This indicates the lifting speed of a single-depth stacker crane; This indicates the speed of the telescopic fork of a single-depth stacker crane; This indicates the travel speed of the double-deep stacker; This indicates the lifting speed of the double-deep stacker; Indicates the speed of the telescopic fork of a double-deep stacker; This represents the objective function value indicating the waste of space in the width direction of the factory building; This represents the value of the overall cost objective function; This represents the objective function value for the overall operating efficiency of stacker cranes in all aisles; This indicates the width of the factory building that can be used for stacker crane automated warehouse layout; Indicates the width of a single-depth tunnel; Indicates the width of a double-deep tunnel; Indicates the width of a semi-double-deep tunnel; This indicates the back-to-back spacing between shelves in adjacent aisles; Indicates the space required for the back pull of the shelf; This indicates the speed parameter combination of a single-depth stacker crane. The corresponding equipment cost; This indicates the speed parameter combination of the double-deep stacker crane. The corresponding equipment cost; This indicates the cost per storage location; This indicates the total number of storage locations in the plan; This indicates the speed parameter combination of a single-depth stacker crane. Total length occupied by a single row of shelves Total height occupied by single-row shelving Operational efficiency under certain conditions; This indicates the speed parameter combination of the double-deep stacker crane. Total length occupied by a single row of shelves Total height occupied by single-row shelving Operational efficiency under certain conditions; This indicates the operational efficiency corresponding to a semi-double-depth tunnel. Indicates the efficiency of project requirements; Indicates the number of storage locations in a single storage compartment; This represents the optimal number of columns for a single-depth shelving unit obtained from the first level of optimization. This represents the optimal number of single-depth shelves obtained from the first level of optimization; This represents the optimal number of columns for the double-deep shelving obtained from the first level of optimization; This represents the optimal number of layers for the double-deep shelving obtained from the first layer of optimization; Indicates the minimum number of storage locations required for the project; This indicates the total number of cargo units corresponding to a single-depth aisle; This indicates the total number of cargo units corresponding to a double-deep aisle; This indicates the total number of cargo units corresponding to a semi-double-depth aisle; This indicates the distance between materials on the loading platform and materials on the shelves in the aisle; Indicates the length of a single stored material; This indicates the distance between the first and second deep items in a double-deep shelving unit; This indicates the total length occupied by a single-depth rack or a single row of racks; This indicates the total height occupied by a single-depth rack or a single row of racks; This indicates the total length occupied by a single row of double-deep shelving; This indicates the total height occupied by a single row of double-deep shelving.

5. The method for generating the layout of a stacker crane automated warehouse with engineering constraint coupling according to claim 4, characterized in that: Step S4 uses an improved genetic algorithm with adapted mixed decision variables to solve the problem. The improved genetic algorithm includes: The number of roadways is encoded using integers; Discrete index encoding is used for the stacker crane speed parameter combinations; The feasible templates generated in step S3 are encoded using template indexing. During fitness evaluation, both multi-objective function values ​​and constraint violation rates are calculated simultaneously. Individuals are screened using a non-dominated ranking system prioritizing constraint dominance and crowding calculation. Tournament selection was used in the selection process; During the crossover process, integer-coded genes are exchanged, discrete-index-coded genes are exchanged within the directory index range, and template-index-coded genes are exchanged within the same rule category. During the mutation process, random integer mutations are performed on integer-coded genes, random index mutations within the directory are performed on discrete-index-coded genes, and template-index-coded genes are subjected to similar template-neighborhood mutations.

6. The method for generating the layout of a stacker crane automated warehouse with engineering constraint coupling according to claim 1, characterized in that: The local installation conflict identification in step S4 includes at least one or more of the following: Conflicts include: end interference, insufficient clearance, misalignment of installation holes, mismatch of rules between top and non-top levels, mismatch of rules between first and second depth levels, inconsistency of equipment operating boundaries, and mismatch of position templates. When a minor conflict is detected, the current layer is repaired based on the reverse correction rule sub-library for the number of lanes, template index, or velocity parameter index. When a moderate to severe conflict is identified, the reverse correction rule sub-library is used to backtrack to step S3 to correct the number of shelf layers, shelf columns, top-level template, deep-level template, or location template, and then the corrected template is sent back to step S4 for optimization evaluation.

7. The method for generating the layout of a stacker crane automated warehouse with engineering constraint coupling according to claim 1, characterized in that: The installation-level parameter package formed in step S5 includes at least the following parameters: storage unit size parameters, number of rack layers and columns parameters, aisle centerline and aisle width parameters, stacker crane model and speed parameters, end buffer zone boundary parameters, back-to-back spacing parameters, back-pull space parameters, mounting hole sequence parameters, top clearance parameters, bottom clearance parameters, fire clearance parameters, and bill of materials parameters.

8. The method for generating the layout of a stacker crane automated warehouse with engineering constraint coupling according to claim 1 or 7, characterized in that: The consistency verification in step S6 includes at least the following: consistency verification between layout parameters and equipment selection parameters; consistency verification between rack height parameters and installation hole parameters; consistency verification between aisle width parameters and stacker crane operating boundary parameters; consistency verification between clearance parameters and safety boundary parameters; consistency verification between drawing output parameters and bill of materials parameters; and consistency verification between template rule labels and final delivery parameters. After all consistency checks pass, the drawing generation system outputs the warehouse unit diagram, automated warehouse floor plan, automated warehouse elevation layout, stacker crane parameter table, installation hole location table, clearance check table, and equipment bill of materials table.

Images