Large-scale tradeoff space parameter data processing method, device and equipment

Abstract

The invention provides a large-scale trade-off space parameter data processing method, device and equipment, and relates to the technical field of strategy optimization, and the method comprises the steps: S1, determining a design variable of a design scheme, and calculating trade-off space data; s2, carrying out classified storage on the tradeoff space data and the process parameter data; s3, generating an initial trade-off space according to the trade-off space data; s4, generating a trade-off space parameter browser according to the initial trade-off space; s5, calculating Pareto frontier data of the target balance space; s6, adjusting the adjustment parameters of the target balance space; s7, redrawing the target tradeoff space according to the adjustment parameters; and S8, carrying out labeling, weight calculation, manual operation and set operation on the design scheme to obtain a final weighing scheme. According to the invention, the scale of the trade-off space is reduced by adjusting the parameter constraint through the trade-off space parameter browser, and the redrawing and trade-off efficiency of the large-scale trade-off space is rapidly improved.

Description

Technical Field

[0001] The present application relates to the field of strategy optimization technology, and in particular to a method, device and equipment for processing large-scale trade-off space parameter data. Background Art

[0002] In the solution of the system, the engineering problem needs to be analyzed and a problem definition is generated to cover all aspects of the engineering problem. In particular, the problem definition should specify the different design criteria that any given system architecture should meet, including design constraints and design goals. The problem definition is input into the system that generates solutions, and the system automatically generates many solution options that meet the design criteria, which constitute the trade-off space. However, there are some problems with the current approach. First, for any given problem definition, the system may generate a large number of solution options, resulting in a large trade-off space and the need to process a large amount of data. Usually, people will ignore a large part of the generated solution options, but this may cause many potential "good" solutions to be ignored because these solutions may be found to meet the criteria after proper analysis and consideration. Second, different people often have different opinions on whether a given solution meets a certain criterion. This difference in opinion reduces the efficiency of multiple people jointly analyzing the numerous solution options in the trade-off space. To solve this problem, people usually only consider a subset of solution options that meet the criteria held by all participants, but this may also cause many potential "good" architectures to be ignored and not considered.

[0003] In summary, how to propose a more effective method to deal with large-scale trade-off spaces, effectively reduce the number of options to be weighed by reasonably setting parameters, and obtain an effective final set of solutions to better optimize problem solutions and meet various design standards and requirements has become an urgent problem to be solved. Summary of the invention

[0004] In order to overcome the deficiencies in the background technology, the present invention discloses a large-scale trade-off spatial parameter data processing method, device and equipment.

[0005] In a first aspect, the present invention provides a large-scale trade-off spatial parameter data processing method, comprising the following steps:

[0006] S1. Determine the design variables of the design scheme and calculate the trade-off space data;

[0007] S2, classify and store the weight space data and process parameter data;

[0008] S3, generating an initial trade-off space according to the trade-off space data;

[0009] S4. Generate a trade-off space parameter browser based on the initial trade-off space. The user adjusts the initial trade-off space through the trade-off space browser to obtain the target trade-off space. The parameter browser can interactively and visually adjust the design parameters and the results of interest, support users to customize coordinate axes, color classification, set constraints, and intuitively display the relationship between parameters through data link lines.

[0010] S5. Calculate the Pareto frontier data of the target trade-off space;

[0011] S6, adjusting the adjustment parameters of the target trade-off space;

[0012] S7, redrawing the target trade-off space according to the adjustment parameters;

[0013] S8. Label the design solutions, calculate weights, and perform manual operations and set operations to obtain the final trade-off solution.

[0014] Specifically, step S1 includes:

[0015] S11, determining design variables, and constructing a calculation model according to the design variables and the results of interest; the number of the design variables is M, and the results of interest include N result factors, where M and N are both positive integers greater than 1;

[0016] S12, collecting data related to each design scheme and design variables, and inputting them into the calculation model to obtain the results of interest corresponding to each design scheme;

[0017] S13, selecting two most representative and relevant first result factors and second result factors from the results of interest as the horizontal and vertical coordinates, and determining the measurement unit and value range of the coordinate axis according to the results of interest;

[0018] S14. Calculate the trade-off space data according to the calculation model corresponding to the design variables and the result factors.

[0019] Specifically, step S2 is as follows: marking the values ​​of the first result factor and the second result factor corresponding to each design solution as a point on the corresponding coordinate axis, and the area formed by the coordinate points corresponding to all design solutions is the initial trade-off space.

[0020] Specifically, the parameter browser in step S4 specifically includes:

[0021] The parameter browser displays the design parameters of each solution and the parameters of the results of interest. The user can select any parameter as the X-axis and any parameter as the Y-axis to construct different views of the trade-off space;

[0022] The parameter browser includes a coloring option, and the user can color different types of design parameters or results of interest with different colors, or color design parameters or results of interest with different interval values ​​in a graded manner;

[0023] The parameter browser includes a constraint window setting, and the user can set the constraint window for each parameter. Once the setting is completed, the trade-off space will automatically delete the data that does not meet the constraint conditions and only display the data points that meet the conditions;

[0024] The parameter browser provides parameter connection and visualization settings. Parameters are connected by lines to represent a specific parameter combination of a point in the trade-off space.

[0025] Specifically, step S5 is as follows: using different genetic algorithms to calculate the Pareto frontier data of target trade-off spaces of different dimensions; if the dimension of the target trade-off space is less than or equal to 3, using the NSGA-II algorithm; if the dimension of the target trade-off space is greater than 3, using the NSGA-III algorithm.

[0026] Specifically, step S6 is as follows: obtaining the target touch parameters of the trade-off space display area, and adjusting the trade-off space adjustment parameters through touch; there is a mapping relationship between the touch parameters and the trade-off space adjustment parameters, and when the touch parameters change, the target trade-off space and the trade-off space adjustment parameters change synchronously.

[0027] Specifically, step S7 further includes: when a GPU is detected, redrawing is accelerated by parallel computing of the GPU.

[0028] Specifically, step S8 specifically includes:

[0029] S81. Label the design solutions individually or by category, and then assign weights based on the labels;

[0030] S82, manually screening out design solutions with weights greater than or equal to a specific threshold to form a solution set;

[0031] S83. Further screening is performed in the solution set through set operations to determine the final trade-off solution.

[0032] In a second aspect, the present invention further provides a large-scale trade-off space parameter data processing device, comprising the following units:

[0033] A trade-off space data calculation unit, used for determining design variables of a design solution and calculating trade-off space data;

[0034] A parameter storage unit, used for classifying and storing the weight space data and process parameter data;

[0035] A trade-off space generating unit, used for generating an initial trade-off space according to the trade-off space data;

[0036] A parameter browser generating unit, used to generate a trade-off space parameter browser according to the initial trade-off space, and the user adjusts the initial trade-off space through the trade-off space browser to obtain a target trade-off space;

[0037] A Pareto front data calculation unit, used to calculate the Pareto front data of the target trade-off space;

[0038] A tuning parameter adjustment unit, used to adjust the tuning parameters of the target trade-off space;

[0039] A redrawing unit, used for redrawing the target trade-off space according to the adjustment parameters;

[0040] The annotation unit is used to annotate the design solutions, calculate weights, and perform manual operations and set operations to obtain the final trade-off solution.

[0041] In a third aspect, the present invention also provides an electronic device comprising a processor, a memory, a communication interface, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the processor, and the program comprises instructions for executing the steps of any one of the methods for large-scale trade-off spatial parameter data processing.

[0042] The present invention provides a large-scale trade-off space parameter data processing method, comprising the following steps: S1, determining the design variables of the design scheme and calculating the trade-off space data; S2, classifying and storing the trade-off space data and process parameter data; S3, generating an initial trade-off space according to the trade-off space data; S4, generating a trade-off space parameter browser according to the initial trade-off space; S5, calculating the Pareto front data of the target trade-off space; S6, adjusting the adjustment parameters of the target trade-off space; S7, redrawing the target trade-off space according to the adjustment parameters; S8, marking the design scheme, calculating the weights, and manually operating and performing set operations to obtain the final trade-off scheme. The present invention reduces the scale of the trade-off space by adjusting the parameter constraints through the trade-off space parameter browser, and quickly improves the redrawing and trade-off efficiency of the large-scale trade-off space.

[0043] Furthermore, by allocating intensive computing tasks to multiple threads on the GPU for parallel execution, the present invention can generate a large number of computing results in a short period of time, quickly build and adjust the trade-off space, and provide users with more timely and accurate decision support.

[0044] In addition, the present invention can flexibly draw and observe the trade-off relationship of different design variables in the parameter browser by adjusting parameters and coloring rules, which is clearer and more intuitive.

[0045] In addition, the present invention selects the most appropriate genetic algorithm according to the characteristics of the problem, thereby improving the solution efficiency and the accuracy of the Pareto frontier, ensuring that the diversity of solutions is maintained in low-dimensional problems, and obtaining a more evenly distributed solution set in high-dimensional problems. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0047] Figure 1 It is a flowchart of a method for processing large-scale weighed spatial parameter data provided according to an embodiment of the present invention;

[0048] Figure 2 It is a schematic diagram of a trade-off space with cost as the X-axis and utility as the Y-axis provided according to an embodiment of the present application;

[0049] Figure 3 It is a schematic diagram of the Pareto front of different levels of the trade-off space provided in the embodiment of the present application;

[0050] Figure 4 is a schematic diagram of a trade-off space parameter browser provided according to an embodiment of the present application;

[0051] Figure 5 is a schematic diagram of a large-scale trade-off spatial parameter data processing device provided according to an embodiment of the present invention;

[0052] Figure 6 is a schematic diagram of a large-scale trade-off spatial parameter data processing device provided according to an embodiment of the present invention;

[0053] Among them, 501, the category of the parameters of the trade-off space; 502, the data axis of a specific parameter of the trade-off space; 503, the data scale of the parameter axis; 504, the label of a specific parameter of the trade-off space; 505, the X-axis and Y-axis buttons; 506, the color and size of the trade-off point in the trade-off space; 507, the parameter adjustment box; 508, the data link line. DETAILED DESCRIPTION

[0054] The present invention can be explained in detail through the following embodiments. The purpose of disclosing the present invention is to protect all technical improvements within the scope of the present invention. In the description of the present invention, it should be understood that if there are terms such as "upper", "lower", "front", "back", "left", "right", etc. indicating directions or positional relationships, they only correspond to the drawings of the present application for the convenience of describing the present invention, and do not indicate or imply that the device or element referred to must have a specific direction.

[0055] Embodiment 1

[0056] refer to Figure 1 This embodiment provides a large-scale trade-off space parameter data processing method, including:

[0057] S1. Determine the design variables of the design scheme and calculate the trade-off space data;

[0058] S11, determining design variables, and constructing a calculation model according to the design variables and the results of interest; the number of the design variables is M, and the results of interest include N result factors, where M and N are both positive integers greater than 1;

[0059] Identify design variables: Identify the various design variables relevant to the system under study that may affect different aspects of the results.

[0060] Select an appropriate computational model: Based on the problem definition and relevant data, select a computational model that can map the design variables to the relevant results of interest. There are various methods for building models, such as mathematical formulas, statistical analysis, simulation, etc.

[0061] For example, to design a new product, design variables may include material type, product size, color, cost, user experience characteristics, etc. Different combinations and choices of these design factors will affect the performance, cost, efficiency, market appeal and other aspects of the product.

[0062] After clarifying the design variables, the next step is to find a computational model that can link these variables with the expected results of interest (such as performance, cost, efficiency, etc.) and form a trade-off space after calculation.

[0063] The construction of the computational model can be based on a variety of methods such as mathematical formulas (such as optimization algorithms, statistical models), statistical analysis (such as regression analysis), and simulation (such as Monte Carlo simulation).

[0064] In this embodiment, the number of design variables is at least M, where M is a positive integer greater than 1, and the corresponding results of interest also include N result factors, where N is a positive integer greater than 1;

[0065] S12, collecting data related to each design scheme and design variables, and inputting them into the calculation model to obtain the results of interest corresponding to each design scheme;

[0066] Collect relevant data: To apply the selected computational model, it is necessary to collect data related to the design variables and the results of interest. The accuracy and completeness of the data are critical to generating a reliable trade-off space.

[0067] Run model mapping: Input the collected design variable data into the calculation model, and obtain the result value of each design scheme corresponding to the result of interest through calculation and analysis.

[0068] Specifically, all combinations of design variable values ​​are traversed, each combination forms a solution, and the design variable data of each combination is input into the calculation model for calculation and analysis to obtain the corresponding results;

[0069] S13, selecting two most representative and relevant first result factors and second result factors from the results of interest as the horizontal and vertical coordinates, and determining the measurement unit and value range of the coordinate axis according to the results of interest;

[0070] Select coordinate axes: Select the two most representative and relevant outcome factors from the outcomes of interest as the horizontal axis (X-axis) and the vertical axis (Y-axis), for example, in the trade-off space, the first outcome factor (cost) is used as the X-axis and the second outcome factor (utility) is used as the Y-axis.

[0071] Determine the units of measurement: Determine the appropriate units of measurement for the selected axis factors to ensure accurate quantification and representation.

[0072] Set value range: Define the possible value range of each axis factor to cover all possible situations.

[0073] S14. Calculate the trade-off space data according to the calculation model corresponding to the design variables and the result factors.

[0074] Select two most representative and relevant result factors from the results of interest, obtain the calculation model corresponding to the first result factor and the second result factor, and calculate the corresponding coordinate point data according to the design variables and the calculation model, that is, the trade-off space data;

[0075] S15. After calculating the trade-off space data, it is necessary to analyze the trade-off relationship, consider the constraints, and make optimizations and decisions.

[0076] After obtaining the weighted spatial data, the following steps are required:

[0077] Analyze trade-offs: Observe the distribution of coordinate points in the trade-off space data, analyze the trade-offs between different factors, and identify different situations such as high cost and high utility, or low cost and low utility.

[0078] Consider constraints: Consider the actual situation and take into account various constraints such as budget limitations, technical requirements, resource availability, etc., which may limit the range of feasible design solutions in the trade-off space.

[0079] Optimization and decision-making: Optimization and decision-making are performed based on the analysis of trade-off spatial data and the consideration of constraints. Design solutions that optimize key results while satisfying constraints can be sought, or appropriate solutions can be selected based on specific goals and preferences.

[0080] In practice, specific steps and methods vary depending on the research field, the complexity of the problem, and the available data and tools. For example, in ecosystem service trade-off studies, the quantification and analysis of multiple services will be involved; in engineering design or economic decision-making, more specific technical and market factors may need to be considered. At the same time, in order to improve the accuracy and reliability of the trade-off space, it may be necessary to continuously improve the model, improve data collection, and conduct sensitivity analysis.

[0081] S2, classify and store the weight space data and process parameter data;

[0082] This embodiment aims to effectively store various process parameters involved in the process of generating a large-scale trade-off space. The process parameters include: design variable related parameters, model related parameters, collected data related parameters, operation model mapping related parameters, coordinate axis related parameters, measurement unit related parameters, value range related parameters, coordinate point related parameters, analysis trade-off relationship related parameters, constraint condition related parameters and optimization and decision related parameters.

[0083] Through appropriate data structures and storage methods, subsequent queries, analysis, and decision-making are facilitated to support in-depth research and application of the trade-off space. Establish a database or data file and design it as follows:

[0084] Determine the storage of design variable related parameters: Create a data table to record various design variables, such as the name, type, value range, etc. of the design variable.

[0085] Storage of model-related parameters: set up special fields to record the name, characteristics, construction method and other information of the selected model;

[0086] Storage of collected data related parameters: Construct data tables to save collected data related to design variables and results to ensure data accuracy and completeness.

[0087] Storage of parameters related to running model mapping: recording intermediate data and final result values ​​during model calculation and analysis.

[0088] Storage of selected axis related parameters: Store the name and characteristics of the selected axis factors.

[0089] Determine the storage of measurement unit related parameters: record the specific measurement units determined for the coordinate axis factors.

[0090] Set the storage of parameters related to the value range: save the value range information of each coordinate axis factor.

[0091] Storage of coordinate point related parameters: record the coordinate point data corresponding to each design scheme on the coordinate axis.

[0092] Storage of trade-off space data: Use appropriate data structures to store the numerous coordinate points that make up the trade-off space.

[0093] Storage of parameters related to trade-off analysis: Save the analysis results of the trade-off relationship, such as descriptions and summaries of different situations.

[0094] Consider the storage of constraint-related parameters: record the specific content and restriction range of various constraints.

[0095] Storage of optimization and decision-making related parameters: Storage of relevant data in the optimization and decision-making process and the final selected solution.

[0096] Through the above comprehensive and organized storage design, the key information in the process of generating the trade-off space can be systematically saved, providing a solid data foundation and support for subsequent work.

[0097] S3, generating an initial trade-off space according to the trade-off space data;

[0098] Draw coordinate points: mark the value of the coordinate axis factor corresponding to each design plan as a point on the corresponding coordinate axis.

[0099] Generate initial trade-off space: The area formed by the coordinate points of many design solutions is the initial trade-off space.

[0100] The trade-off space is fully displayed with cost as the X-axis and utility as the Y-axis, and each point in the initial trade-off space represents a specific solution. The embodiment of the trade-off space can be displayed in an electronic device, and the trade-off space can be enlarged or reduced as needed. When clicking on a parameter point in the trade-off space, the detailed parameters of the solution point can be viewed in the parameter browser.

[0101] This embodiment provides a specific embodiment method for generating a vehicle battery architecture selection trade-off space, as follows:

[0102] Determine the design variables (design variable definition):

[0103] 1) Type of battery cell: Assume the value is 1 (lithium iron phosphate) or 2 (ternary lithium)

[0104] 2) Number of battery cells (Number): value range [100,500], step size is 10

[0105] 3) Battery module arrangement (Layout): Assume the value is 1 (series), 2 (parallel), 3 (combination of series and parallel)

[0106] 4) Cooling method of battery pack (Cooling): Assume the value is 1 (air cooling) or 2 (liquid cooling)

[0107] 5) Complexity of battery management system (BMS_Complexity): value range [1,5], 1 represents the simplest and 5 represents the most complex

[0108] 6) Battery energy density (Energy_Density): value range [100,300] Wh / kg, step size is 10

[0109] 7) Battery charge and discharge rate (Charge_Discharge_Rate): value range [1C, 5C], step size is 0.5C

[0110] 8) Battery cycle life (Cycle_Life): value range [1000,5000] times, step length is 100 times

[0111] 9) Battery safety protection level (Safety_Level): value range [1,5], 1 represents the lowest level, 5 represents the highest level

[0112] 10) Battery packaging material (Packaging_Material): Assume the value is 1 (metal), 2 (plastic)

[0113] Cost Calculation

[0114] Cost = battery cell cost + module assembly cost + cooling system cost + BMS cost + packaging material cost

[0115] Assume that the battery cell cost (Cost_Battery_Cell) = number of cells × unit price (determined by cell type)

[0116] Module assembly cost (Cost_Module_Assembly) = Number of modules × Module assembly unit price

[0117] Cost of cooling system (Cost_Cooling_System) = unit price of cooling method × correlation coefficient (determined according to cooling method)

[0118] BMS cost (Cost_BMS) = BMS complexity × BMS unit price

[0119] Cost of packaging materials (Cost_Packaging) = unit price of packaging materials × correlation coefficient (determined according to packaging materials)

[0120] Total cost formula:

[0121] Cost=Number×Unit_Price_Type×Coefficient_Type+Module_Number×Assembly_Price+Cooling_Price×Coefficient_Cooling+BMS_Complexity×BMS_Price+Packaging_Price×Coefficient_Packaging

[0122] Utility Computing

[0123] Utility = driving range utility + charging time utility + battery life utility + safety utility

[0124] Utility_Range = Actual cruising range / Target cruising range

[0125] Assume that the actual driving range is related to the battery energy density and total capacity:

[0126] Actual driving range (Range_Actual) = total battery capacity (calculated based on the number of cells and energy density) × efficiency coefficient

[0127] Utility_Charge_Time = 1 / actual charging time

[0128] Assume that the actual charging time is related to the charge and discharge rate:

[0129] Actual charging time (Charge_Time_Actual) = battery capacity / charge and discharge rate × charging efficiency

[0130] Battery life utility (Utility_Cycle_Life) = actual cycle life / target cycle life

[0131] Safety utility (Utility_Safety) = Safety protection level / Maximum safety level

[0132] Total utility formula:

[0133] Utility=(Range_Actual / Target_Range)+(1 / Charge_Time_Actual)+(Cycle_Life_Actual / Target_Cycle_Life)+(Safety_Level / Max_Safety_Level)

[0134] Calculate the trade-off space data: traverse all the value combinations of the design variables, and each combination forms a solution;

[0135] For each combination, the cost and utility values ​​are calculated according to the cost and utility formulas above;

[0136] Select cost and utility as the outcome factors of interest, with the first outcome factor (cost) as the X-axis and the second outcome factor (utility) as the Y-axis;

[0137] The calculated cost is used as the X-axis coordinate and the utility is used as the Y-axis coordinate, and they are plotted in a two-dimensional coordinate system;

[0138] Repeat the above steps until the complete coordinate points are generated. Through the above detailed process, the trade-off space of the selection of automotive battery architecture can be generated, thereby intuitively showing the trade-off relationship between cost and utility of different design schemes. Figure 2 shown.

[0139] For the design variables involved in this embodiment, the type of battery cell has 2 values; the number of battery cells has a step size of 10, a value range of [100,500], and a total of (500-100) / 10+1=41 values; the arrangement of battery modules has 3 values; the cooling method of the battery pack has 2 values; the complexity of the battery management system has a value range of [1,5], and a total of 5 values; the energy density of the battery has a step size of 10, and a value range of [100,3 00], there are (300-100) / 10+1=21 values; the charge and discharge rate of the battery, the step size is 0.5C, the value range is [1C, 5C], there are (5-1) / 0.5+1=9 values; the cycle life of the battery, the step size is 100 times, the value range is [1000, 5000], there are (5000-1000) / 100+1=41 values; the safety protection level of the battery, there are 5 values; the packaging material of the battery, there are 2 values. The total number of options in the trade-off space is 7879200, which also means that there are 7879200 solution points in the trade-off space. It is necessary to find a satisfactory solution among these solutions for decision makers to make decisions. However, it is extremely difficult to make trade-offs among many solutions.

[0140] S4, generating a trade-off space parameter browser according to the initial trade-off space, and the user adjusts the initial trade-off space through the trade-off space browser to obtain the target trade-off space;

[0141] The parameter browser can interactively visualize and adjust design parameters and results of interest, support users to customize coordinate axes, color classification, set constraints, and intuitively display the relationship between parameters through data link lines;

[0142] The parameter browser displays the design parameters of each solution and the parameters of the results of interest. The user can select any parameter as the X-axis and any parameter as the Y-axis to construct different views of the trade-off space;

[0143] The parameter browser includes a coloring option, and the user can color different types of design parameters or results of interest with different colors, or color design parameters or results of interest with different interval values ​​in a graded manner;

[0144] The parameter browser includes a constraint window setting, and the user can set the constraint window for each parameter. Once the setting is completed, the trade-off space will automatically delete the data that does not meet the constraint conditions and only display the data points that meet the conditions;

[0145] The parameter browser provides parameter connection and visualization settings. Parameters are connected by lines to represent a specific parameter combination of a point in the trade-off space.

[0146] Figure 3 An example of a trade-off space parameter browser is given. 501-508 represent the main components of the trade-off space parameter browser, where 501 represents the category of the generated trade-off space parameter, such as capability attributes, design variables, etc.; 502 represents the data axis of a specific parameter in the trade-off space; 503 represents the data scale of the parameter axis; 504 represents the label of a specific parameter in the trade-off space, corresponding to the data of the parameter axis; 505 represents the X-axis and Y-axis buttons, indicating that the parameter should be displayed on the X-axis or the Y-axis; 506 represents the color and size table of the trade-off point in the trade-off space. 507 is a parameter adjustment box, which is used to adjust the parameter range of the corresponding constraint parameter. Multiple parameters can be adjusted at the same time. After the adjustment, the solution points in the trade-off space are displayed according to the parameter constraints. 508 represents a data link line, which links two different parameter axes at the actual parameter value. When there is no parameter adjustment box set, all data link lines can be displayed. After the parameter adjustment box is set, only the data connection lines with the parameter adjustment box are displayed, and the links outside the parameter range are displayed in a lighter color or not displayed at all.

[0147] By defining multiple parameters, calculating costs and utilities, and using parameter browsers for visualization and interactive operations, users can clearly see the relationship between costs and utilities of different design solutions, and adjust and filter them according to their needs to find the optimal solution. This system includes a parameter browser module, a trade-off space generation module, a coloring and classification processing module, a constraint processing module, and a parameter connection and visualization module.

[0148] The trade-off space parameter browser is automatically generated according to demand or generated according to user needs; the trade-off space parameter browser can be manually adjusted according to user needs; when the parameters in the trade-off space parameter browser change, the trade-off space will change accordingly with the change of the parameter browser. If the parameter browser contains data connections, the data connections will change accordingly with the change of parameter constraints. Those within the constraint range will be displayed, those outside the constraint range will not be displayed, or will be displayed in a lighter color.

[0149] Taking the battery selection embodiment as an example, the user opens the parameter browser, which contains design variables such as the type of battery cells, the number of battery cells, the arrangement of battery modules, the cooling method of the battery pack, the complexity of the battery management system, the energy density of the battery, the charge and discharge rate of the battery, the cycle life of the battery, the safety protection level of the battery, the packaging material of the battery, and parameter options such as cost and utility. The user can freely select any parameter as the X-axis and any parameter as the Y-axis to construct different views of the trade-off space.

[0150] Coloring options: For example, users can color different battery cell types (such as blue for lithium iron phosphate and red for ternary lithium), or color data points in different cost ranges (such as green for costs below 5,000 yuan, yellow for 5,000-10,000 yuan, and orange for costs above 10,000 yuan) in different colors to quickly distinguish and classify data.

[0151] Constraint window setting: Users can set constraint windows for each parameter. For example, set the number of battery cells to [200,400] and the complexity constraint window of the battery management system to [2,4]. Once the setting is completed, the trade-off space will automatically delete data that does not meet the constraint conditions and only display data points that meet the conditions.

[0152] Parameter connection and visualization: Parameters are connected by lines to represent the specific parameter combination of a point in the trade-off space. For example, when the number of battery cells is selected as the X-axis and the utility is selected as the Y-axis, a point (300, utility value 0.8) is connected by lines to the corresponding parameters such as battery cell type (ternary lithium) and battery module arrangement (series). Users can intuitively see the corresponding status of other parameters as the number of battery cells changes.

[0153] Application scenario example: Suppose an automobile manufacturer is selecting a new battery architecture for electric vehicles. The user first selects the X-axis as cost, the Y-axis as utility, and the coloring basis as the energy density of the battery in the parameter browser. Then set the constraint window, such as the cost does not exceed 8,000 yuan and the utility is not less than 0.7. After the system generates the trade-off space, the user can clearly see the distribution of data points that meet the conditions. Through line connections, users can intuitively understand the specific battery architecture parameters corresponding to each point, such as the type of battery cell is lithium iron phosphate, the number is 350, and the module arrangement is parallel. In the end, users can quickly determine the optimal battery architecture solution that meets cost and utility requirements.

[0154] Through the above embodiments, efficient, intuitive and flexible trade-off space adjustment and decision support can be provided through the ability to select.

[0155] S5. Calculate the Pareto frontier data of the target trade-off space;

[0156] Different genetic algorithms are used to calculate the Pareto frontier data of target trade-off spaces of different dimensions. If the dimension of the target trade-off space is less than or equal to 3, the NSGA-II algorithm is used; if the dimension of the target trade-off space is greater than 3, the NSGA-III algorithm is used.

[0157] The same part of NSGA (non-dominated sorting genetic algorithm), NSGA-II and NSGA-III is that they all adopt the idea of ​​non-dominated sorting, and classify the individuals of the population into different non-dominated frontiers through fast non-dominated sorting.

[0158] Their differences are mainly reflected in the environmental selection stage, that is, how to select individuals from the non-dominated layer to enter the next generation of population. NSGA-II uses crowding distance to maintain the diversity of the population. After determining the individuals in each non-dominated layer, their crowding distance in the target space is calculated. Individuals with large crowding distances are more likely to be selected to enter the next generation of population. NSGA-III uses reference points to maintain the diversity of the population. The specific steps are as follows:

[0159] 1. Generation of reference points: The reference points can be placed on a standardized hyperplane by using the boundary intersection weighting method. For a problem with M targets, if we consider dividing it into p parts along each target direction, the total number of reference points H can be calculated by the formula To avoid a sharp increase in the number of reference points when the target dimension is large, a two-layer reference point generation method can be used, adding an inner layer to ensure the generation of intermediate points without increasing the number of reference points too much.

[0160] 2. Standardized target space: First, obtain the minimum value of all individuals in the population St on each dimension of the target to form the ideal point of the current population. Then, convert the target values ​​and ideal points of all individuals with reference to the ideal point to obtain a temporary standardized target value. Then calculate the extreme points on each dimension of the target axis. These extreme points form an M-1-dimensional linear hyperplane, from which the intercepts in each target direction can be calculated. Finally, use the intercept and the temporary standardized target value to calculate the true standardized target value: where a i is the intercept in each target direction. The method of generating extreme points is to use the weight vector w as the achievement scalarizing function (ASF function) in the direction of the coordinate axis. For example, to calculate the extreme point on the target axis of the i-th dimension, the target direction needs to be fixed so that the direction vector ω i =1, the weights of other directions are set to 10e-6 (infinitely close to 0), and then the ASF function As the objective function, it is minimized, where x is the individual, ω is the weight vector, and f m (x) is the individual f m The value of (x) on the mth target, is the ideal point of the mth target.

[0161] 3. Environmental selection: Assumption F l is the last non-dominated layer that is larger than the population size N for the first time. The purpose is to find F l N-|S t | optimal individuals, merged into the next generation population S t+1 The specific operation is to calculate the degree of association between individuals in the population and the reference point (such as vertical distance, etc.), and select individuals based on the degree of association.

[0162] In principle, the crowding distance method of NSGA-II can effectively maintain diversity when the dimension of the target optimization problem is low, but as the number of targets increases, its effect may deteriorate and the computational complexity is also high. The reference point method of NSGA-III is more suitable for dealing with high-dimensional target optimization problems. The distribution of reference points guides the algorithm to search for non-dominated solutions with more uniform distribution, and it is relatively more efficient in computation. However, the setting of reference points and the measurement method of the degree of association in NSGA-III may need to be adjusted and optimized according to the specific problem. In practical applications, the system can automatically select which algorithm to use to calculate the Pareto frontier according to the characteristics of the problem. For example, for problems with low target dimensions and high diversity requirements, NSGA-II may be more suitable; while for high-dimensional target optimization problems, NSGA-III usually performs better. Therefore, for different most suitable application scenarios, the system can use NSGA, NSGA-II, NSGA-III or the variation algorithms of these algorithms to obtain the Pareto frontier through manual or automatic matching by the user.

[0163] The following is an example description of the process of generating the above-mentioned automobile battery architecture selection trade-off space:

[0164] In this example of selecting a car battery architecture, we have multiple design variables, each with a different range and method of taking values. For example, the types of battery cells can be lithium iron phosphate or ternary lithium; the number of cells can vary from 100 to 500 in steps of 10, etc.

[0165] When we consider the application of NSGA-II and NSGA-III algorithms, for relatively simple situations, such as when the dimension of the target optimization problem is low, when the target trade-off space has only two main objectives, cost and utility, and when the requirements for solution diversity are high, the NSGA-II algorithm can be used.

[0166] For example, when calculating the crowding distance, for a specific combination of solutions, if its cost is low and its utility is high, and there are fewer similar solutions around it, then its crowding distance is larger and it is more likely to be selected into the next generation of the population, thereby gradually finding better solutions while ensuring a certain degree of diversity.

[0167] However, if the problem becomes more complicated, for example, we need to consider more objectives such as battery recycling efficiency and environmental impact, the dimension of the objective trade-off space becomes higher, forming a high-dimensional objective optimization problem. At this time, the NSGA-III algorithm may be more applicable.

[0168] If the dimension of the target trade-off space is less than or equal to 3, it is judged as a low-dimensional target optimization problem and the NSGA-II algorithm is used; if the dimension of the target trade-off space is greater than 3, it is judged as a high-dimensional target optimization problem and the NSGA-III algorithm is used.

[0169] refer to Figure 4 The figure shows a schematic diagram of the Pareto frontier of different levels in the trade-off space, where Rank represents the Pareto level. The Pareto level is a method used to evaluate the quality of solutions in multi-objective optimization problems. The smaller the level number, the better the solution performs in multi-objective optimization.

[0170] NSGA, NSGA-II and NSGA-III algorithms can all calculate the Pareto Rank. The solutions in the same rank have the same advantages and disadvantages, which is equivalent to stratifying different solutions, similar to product grading. Decision makers can clearly see at which level of the Pareto Rank their decision choices are. Rank is a very critical parameter, and we can use this parameter to make trade-offs. In general, we prefer the solution in Rank1, but if Rank1 is empty under the constraints, then we have to follow the same logic and choose a lower level Rank.

[0171] In NSGA-III, reference points are generated first. Suppose we divide each target direction into multiple parts, and then evaluate the solutions based on these reference points. For example, if a solution has a better distribution relationship with the reference points than expected, it is more likely to be selected. In this way, according to the specific situation of the problem and the target dimension, the appropriate algorithm is automatically selected to handle the trade-offs and screening of a large number of solutions in the selection of automotive battery architecture, so as to find the optimal trade-off between multiple objectives such as cost and utility. Compared with the previous example, this example combined with the selection of automotive battery architecture can more specifically demonstrate the application and selection of NSGA-II and NSGA-III algorithms in actual complex multi-objective optimization problems.

[0172] S6, adjusting the adjustment parameters of the target trade-off space;

[0173] The target touch parameters of the trade-off space display area are obtained, and the trade-off space adjustment parameters are adjusted by touch.

[0174] There is a mapping relationship between the touch parameter and the trade-off space adjustment parameter, and when the touch parameter changes, the target trade-off space and the trade-off space adjustment parameter change synchronously;

[0175] There is a mapping relationship between the touch parameters and the target operation. The corresponding target operation is triggered by touching the parameters, and then the preference parameters are adjusted by touch; the target operations include: upload operation, download operation, update operation, new operation, edit operation, delete operation, and refresh operation.

[0176] Touch parameters are data generated when a user touches the screen. These parameters can be detected by electronic devices and used to understand the user's intentions;

[0177] The electronic device pre-stores the mapping relationship between preset touch parameters and target operations. This means that when the user touches the screen in a specific way (such as long press or fast slide), the system can recognize this touch mode and map it to a specific target operation.

[0178] For example, if a user wants to edit a preference setting, they can do so by sliding or pressing the setting on the screen. After the system detects these touch parameters, it determines that the user's target operation is to edit based on the preset mapping relationship and opens the editing interface.

[0179] For example, if the user performs a long press operation, the system may recognize the corresponding target operation as an edit operation, thereby allowing the user to change the adjustment parameters of the current target trade-off space.

[0180] In an example of a possible trade-off space adjustment parameter, the following steps may be included: obtaining the target touch parameter of the trade-off space display area, adjusting the trade-off space adjustment parameter by touch; determining the target operation corresponding to the target touch parameter according to the preset mapping relationship between the touch parameter and the operation. Wherein, in the embodiment of the present application, the touch parameter may be at least one of the following: touch force, touch time, touch position, touch trajectory, number of touches, touch frequency, touch direction, touch area, etc., which are not limited here. The touch force may be the average touch force when the touch area is within the preset area range, or the touch force may be the maximum touch force when the touch time is greater than the preset duration, etc. The preset area range and the preset duration may be set by the user or by the system default. The touch time may be a time point or a time length, for example, the touch time is the duration when the touch force is greater than the preset touch force, or the touch time may be the time point when the touch area is equal to the specified touch area, etc., wherein the preset touch force and the specified touch area may be set by the user or by the system default.

[0181] In a specific implementation, the mapping relationship between preset touch parameters and operations can be pre-stored in the electronic device, and then, when the user touches the adjustment parameter that needs to be adjusted, the trade-off space adjustment parameter can be adjusted through the touch display screen. The decision preference map and the trade-off space adjustment parameter change synchronously according to the touch content. This step is mainly to improve the efficiency of adjusting parameter setting. Among them, for part or all of the preset content.

[0182] S7, redrawing the target trade-off space according to the adjustment parameters;

[0183] Step S7 also includes: when a GPU is detected, accelerating the redrawing through parallel computing of the GPU;

[0184] According to the value range and number of design variables, the computing task is divided into A thread blocks, each thread block contains B threads, each thread and each GPU thread is responsible for processing the calculation of a design variable combination, and the first result factor and the second result factor values ​​calculated by the GPU are transmitted back to the CPU memory; the results are screened and sorted on the CPU according to the pre-set constraints, so as to achieve the acceleration effect; where A is a positive integer greater than 1, and B is a positive integer greater than 1.

[0185] In the process of calculating and displaying large-scale trade-off spaces, the core principle of using GPUs for calculations is to use the GPU's large-scale parallel processing capabilities to accelerate the calculation of cost and utility values ​​for many combinations of design variables. Traditional CPU calculations are inefficient when processing large-scale data and complex calculations, while GPUs have a large number of computing cores and can handle multiple computing tasks at the same time, thereby greatly improving the calculation speed. By assigning computationally intensive cost and utility calculation tasks to multiple threads on the GPU for parallel execution, a large number of calculation results can be generated in a short period of time, and the trade-off space can be quickly constructed and adjusted, providing users with more timely and accurate decision support.

[0186] To use GPU acceleration to display points in trade-off space, there are several key steps to consider:

[0187] Data transfer: Efficiently transfer point data in the trade-off space from CPU memory to GPU memory. This usually requires the use of specific APIs or libraries to implement fast data copies.

[0188] Parallel computing: Utilize the powerful parallel processing capabilities of the GPU to calculate and process multiple points at the same time. For example, the coordinate calculation, color calculation, lighting effect, etc. of each point can be assigned to different GPU threads for parallel execution.

[0189] Graphics rendering pipeline optimization: Make full use of the GPU's graphics rendering pipeline, such as vertex processing, fragment processing, etc. Improve rendering efficiency by properly setting vertex buffers, index buffers, etc.

[0190] Algorithm optimization: Based on the characteristics of points in the balance space, select an algorithm suitable for GPU parallel computing. For example, use a GPU-based sorting algorithm to sort the points, or use a parallel reduction algorithm to calculate certain statistical information of the points.

[0191] Shader Programming: Write efficient vertex shaders and fragment shaders to process point attributes and effects. Complex calculations can be implemented in shaders, such as calculating color, transparency, etc. based on the position and other attributes of the point.

[0192] Batch processing: Group point data for batch processing to reduce the communication overhead between the GPU and CPU.

[0193] Memory management optimization: Rationally allocate and use GPU memory to avoid memory fragmentation and unnecessary memory copies.

[0194] Take advantage of GPU-specific features: For example, use the GPU's texture mapping to store point-related information, or use the GPU's hardware acceleration features such as geometry shaders, compute shaders, etc.

[0195] Graphics libraries OpenGL and DirectX can be used for debugging and optimization, and the version of the specific graphics library is not limited here.

[0196] If there is a GPU in the electronic device, the GPU acceleration option is automatically enabled in the software, and the specific GPU device in the server is specified to be used. The specific GPU model and brand are not limited here. According to the value range and number of design variables, the computing task is reasonably divided into multiple thread blocks, and each thread block contains several threads. For example, for the variable of the number of battery cells, the corresponding number of threads can be assigned according to its value range and step size. Each GPU thread is responsible for processing a specific combination of design variables. According to the assigned design variable values, the threads call the cost and utility calculation formulas for fast calculation at the same time. Due to the parallel characteristics of the GPU, thousands of threads can be calculated at the same time, greatly improving the computing efficiency. The large number of cost and utility values ​​calculated by the GPU are quickly transferred back to the CPU memory. According to the pre-set constraints, the results are screened and sorted on the CPU using efficient algorithms.

[0197] In the software's visual interface, with cost as the X-axis and utility as the Y-axis, the filtered valid data points are plotted to form a clear and intuitive trade-off space. Users can flexibly adjust the coloring rules through the parameter browser, such as coloring according to the battery cell type, or coloring by classification according to the cost range. Users can also freely select different design variables as the X-axis or Y-axis to observe the trade-off relationship from different angles. Whenever the user modifies the constraints (such as adjusting the cost upper limit, utility lower limit), coloring rules, or coordinate axis selection, the system immediately repeats the steps and quickly updates the display of the trade-off space, allowing users to see the results of the changes in real time, so as to analyze and make decisions more effectively. For example, initially the user observes the trade-off space with cost as the X-axis and utility as the Y-axis, and then decides to modify the coloring rules to distinguish according to the cycle life of the battery and adjust the cost constraint to no more than 8,000 yuan. The system responds quickly, recalculates and updates the trade-off space, clearly showing the data distribution and coloring that meet the new conditions, helping users quickly identify potential high-quality automotive battery architecture selection solutions that meet their latest needs.

[0198] S8. Label the design solutions, calculate weights, and perform manual operations and set operations to obtain the final trade-off solution.

[0199] S81. Label the design solutions individually or by category, and then assign weights based on the labels;

[0200] S82, manually screening out design solutions with weights greater than or equal to a specific threshold to form a solution set;

[0201] S83. Further screening is performed in the solution set through set operations to determine the final trade-off solution.

[0202] Mark the solutions you are interested in. You can mark a single solution, a category of solutions, or all solutions based on constraints, and add or subtract solutions based on your preferences.

[0203] The following is a method that integrates the relevant content, including solution annotation, weight calculation, and finding a satisfactory solution through manual operation and set operation:

[0204] Notation: Let the solution set be S = {S 1 ,S 2 ,S 3 ,…,S n}. Mark a specific solution separately: mark a specific solution directly, for example, mark solution S i The options are “extremely prefer”, “very much prefer”, “strongly prefer”, “somewhat prefer”, “moderately prefer”, “slightly prefer”, “no preference” or “don’t know preference”.

[0205] Label a class of solutions: You can define the characteristics of a class of solutions, classify solutions with these characteristics into a class and label them. Label all solutions based on constraints: Label the entire set of solutions based on certain constraints.

[0206] Weight calculation: Assign a corresponding weight to each preference level, for example, W(S i )=9, strongly prefer W(S i )=7, strongly prefer: W(S i )=5, prefer W(S i )=3, generally prefer W(S i )=2, slightly prefer W(S i )=1.5, no preference W(S i )=1, unknown preference: W(S i )=0.5.

[0207] Manual operation and set operation to find satisfactory solutions, manually filter out solutions with weights greater than or equal to a specific threshold, form set A, and further manually select or exclude from set A according to specific needs. The specific threshold is determined according to specific needs. Perform set operations such as intersection and union on the solutions in set A to obtain the final satisfactory solution set. The set operation method used is determined according to specific needs. For example, suppose there is a solution set S = {S 1 ,S 2 ,S 3 ,S 4}, the labels and weights are, S 1 Very much preferred, W(S 1 )=9,S 2 No preference, W(S 2 )=1,S 3 I don't know the preference, W(S 3 )=0.5,S 4 Based on some constraints being marked as strongly preferred, W(S 4 )=5. Let the threshold be 3, then the set A={S 1 ,S 4}. You can operate on set A according to further requirements, such as if you need to satisfy solution S at the same time 1 and S 4 If the conditions are met, then an intersection operation is performed; if only one of them needs to be met, then a union operation is performed to find the final satisfactory solution, that is, the final trade-off solution. In practical applications, these methods can be flexibly used according to specific circumstances to determine a satisfactory solution.

[0208] In the above embodiment, we adopt a hierarchical approach to divide preferences into different levels and set corresponding weights for each level, so as to more carefully reflect the degree of preference for different solutions. In this way, the decision maker's wishes and preferences can be more accurately reflected in the process of solution labeling and screening. However, please note that this is just a simple example. In actual applications, the preference level and weight settings may need to be further adjusted according to specific circumstances.

[0209] The present embodiment provides a large-scale trade-off space parameter data processing method, comprising the following steps: S1, determining the design variables of the design scheme and calculating the trade-off space data; S2, classifying and storing the trade-off space data and process parameter data; S3, generating an initial trade-off space according to the trade-off space data; S4, generating a trade-off space parameter browser according to the initial trade-off space; S5, calculating the Pareto front data of the target trade-off space; S6, adjusting the adjustment parameters of the target trade-off space; S7, redrawing the target trade-off space according to the adjustment parameters; S8, annotating the design scheme, calculating the weights, and performing manual operations and set operations to obtain the final trade-off scheme. The present embodiment reduces the scale of the trade-off space by adjusting the parameter constraints in the trade-off space parameter browser, and quickly improves the redrawing and trade-off efficiency of large-scale trade-off spaces.

[0210] Furthermore, by allocating intensive computing tasks to multiple threads on the GPU for parallel execution, this embodiment can generate a large number of computing results in a short period of time, quickly build and adjust the trade-off space, and provide users with more timely and accurate decision support.

[0211] In addition, this embodiment can flexibly draw and observe the trade-off relationship of different design variables in the parameter browser by adjusting parameters and coloring rules, which is clearer and more intuitive.

[0212] In addition, this embodiment selects the most appropriate genetic algorithm according to the characteristics of the problem, thereby improving the solution efficiency and the accuracy of the Pareto front, ensuring the diversity of solutions in low-dimensional problems, and obtaining a more evenly distributed solution set in high-dimensional problems.

[0213] Embodiment 2

[0214] refer to Figure 5 This embodiment discloses a large-scale trade-off space parameter data processing device, including the following units:

[0215] A trade-off space data calculation unit, used for determining design variables of a design solution and calculating trade-off space data;

[0216] A parameter storage unit, used for classifying and storing the weight space data and process parameter data;

[0217] A trade-off space generating unit, used for generating an initial trade-off space according to the trade-off space data;

[0218] A parameter browser generating unit, used to generate a trade-off space parameter browser according to the initial trade-off space, and the user adjusts the initial trade-off space through the trade-off space browser to obtain a target trade-off space;

[0219] A Pareto front data calculation unit, used to calculate the Pareto front data of the target trade-off space;

[0220] A tuning parameter adjustment unit, used to adjust the tuning parameters of the target trade-off space;

[0221] A redrawing unit, used for redrawing the target trade-off space according to the adjustment parameters;

[0222] The annotation unit is used to annotate the design solutions, calculate weights, and perform manual operations and set operations to obtain the final trade-off solution.

[0223] A large-scale weighed spatial parameter data processing device provided in an embodiment of the present invention can execute a large-scale weighed spatial parameter data processing method provided in any embodiment of the present invention, and has functional modules and beneficial effects corresponding to the execution method. It is worth noting that in the above-mentioned embodiment of the large-scale weighed spatial parameter data processing device, the various units and modules included are only divided according to functional logic, but are not limited to the above-mentioned division, as long as the corresponding functions can be realized; in addition, the specific names of the functional units are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of the present invention.

[0224] Embodiment 3

[0225] refer to Figure 6 , Figure 6 : is a schematic diagram of the structure of a large-scale weighing spatial parameter data processing device of this embodiment. The large-scale weighing spatial parameter data processing device 20 of this embodiment includes a processor 21, a memory 22, and a computer program stored in the memory 22 and executable on the processor 21. When the processor 21 executes the computer program, the steps in the above-mentioned method embodiment are implemented. Alternatively, when the processor 21 executes the computer program, the functions of each module / unit in the above-mentioned device embodiments are implemented.

[0226] Exemplarily, the computer program may be divided into one or more modules / units, which are stored in the memory 22 and executed by the processor 21 to complete the present invention. The one or more modules / units may be a series of computer program instruction segments capable of completing specific functions, which are used to describe the execution process of the computer program in the large-scale trade-off space parameter data processing device 20. For example, the computer program may be divided into the modules in Embodiment 2. For the specific functions of each module, please refer to the working process of the device described in the above embodiment, which will not be repeated here.

[0227] The large-scale trade-off space parameter data processing device 20 may include, but is not limited to, a processor 21 and a memory 22. Those skilled in the art will appreciate that the schematic diagram is merely an example of the large-scale trade-off space parameter data processing device 20 and does not constitute a limitation on the large-scale trade-off space parameter data processing device 20. The large-scale trade-off space parameter data processing device 20 may include more or fewer components than shown in the diagram, or may combine certain components, or different components. For example, the large-scale trade-off space parameter data processing device 20 may also include input and output devices, network access devices, buses, and the like.

[0228] The processor 21 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc. The processor 21 is the control center of the large-scale weighing space parameter data processing device 20, and uses various interfaces and lines to connect various parts of the entire large-scale weighing space parameter data processing device 20.

[0229] The memory 22 can be used to store the computer program and / or module. The processor 21 realizes various functions of the large-scale trade-off space parameter data processing device 20 by running or executing the computer program and / or module stored in the memory 22 and calling the data stored in the memory 22. The memory 22 can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, an application required for at least one function (such as a sound playback function, an image playback function, etc.), etc.; the data storage area can store data created according to the use of the mobile phone (such as audio data, a phone book, etc.), etc. In addition, the memory 22 can include a high-speed random access memory, and can also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), at least one disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0230] Wherein, if the module / unit integrated in the large-scale trade-off space parameter data processing device 20 is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the present invention implements all or part of the processes in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by the processor 21, the steps of the above-mentioned method embodiments can be implemented. Wherein, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electrical carrier signal, telecommunication signal and software distribution medium, etc. It should be noted that the content contained in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media does not include electrical carrier signals and telecommunication signals.

[0231] It should be noted that the device embodiments described above are merely schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. In addition, in the accompanying drawings of the device embodiments provided by the present invention, the connection relationship between the modules indicates that there is a communication connection between them, which may be specifically implemented as one or more communication buses or signal lines. A person of ordinary skill in the art may understand and implement it without paying any creative effort.

[0232] This specification is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of this specification. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 Processes Multiple processes and / or boxes Figure 1 A device that has the functions specified in one or more boxes.

[0233] These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0234] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0235] The parts of the present invention that are not described in detail are prior art. It is obvious to those skilled in the art that the present invention is not limited to the details of the above-mentioned exemplary embodiments, and the present invention can be implemented in other specific forms without departing from the spirit or basic characteristics of the present invention; therefore, no matter from which point of view, the embodiments should be regarded as exemplary and non-restrictive, and it is intended that all changes that fall within the meaning and scope of equivalent elements are included in the present invention.

Claims

1. A method for processing large-scale trade-off space parameter data, characterized in that: The following steps are described: S1. Determine the design variables of the design scheme and calculate the trade-off space data; S2, classify and store the weight space data and process parameter data; S3, generating an initial trade-off space according to the trade-off space data; S4, generating a trade-off space parameter browser according to the initial trade-off space, and the user adjusts the initial trade-off space through the trade-off space browser to obtain the target trade-off space; The parameter browser can interactively and visually adjust design parameters and focus results, support users to customize coordinate axes, color classification, set constraints, and intuitively display the relationship between parameters through data link lines; S5. Calculate the Pareto frontier data of the target trade-off space; S6, adjusting the adjustment parameters of the target trade-off space; S7, redrawing the target trade-off space according to the adjustment parameters; S8. Label the design solutions, calculate weights, and perform manual operations and set operations to obtain the final trade-off solution.

2. The method according to claim 1, characterized in that Step S1 includes: S11, determining design variables, and constructing a calculation model according to the design variables and the results of interest; the number of the design variables is M, and the results of interest include N result factors, where M and N are both positive integers greater than 1; S12, collecting data related to each design scheme and design variables, and inputting them into the calculation model to obtain the results of interest corresponding to each design scheme; S13, selecting two most representative and relevant first result factors and second result factors from the results of interest as the horizontal and vertical coordinates, and determining the measurement unit and value range of the coordinate axis according to the results of interest; S14. Calculate the trade-off space data according to the calculation model corresponding to the design variables and the result factors.

3. The method according to claim 2, characterized in that Step S2 specifically includes: marking the values ​​of the first result factor and the second result factor corresponding to each design solution as a point on the corresponding coordinate axis, and the area formed by the coordinate points corresponding to all the design solutions is the initial trade-off space.

4. The method according to claim 1, characterized in that: The parameter browser in step S4 specifically includes: The parameter browser displays the design parameters of each solution and the parameters of the results of interest. The user can select any parameter as the X-axis and any parameter as the Y-axis to construct different views of the trade-off space; The parameter browser includes a coloring option, and the user can color different types of design parameters or results of interest with different colors, or color design parameters or results of interest with different interval values ​​in a graded manner; The parameter browser includes a constraint window setting, and the user can set the constraint window for each parameter. Once the setting is completed, the trade-off space will automatically delete the data that does not meet the constraint conditions and only display the data points that meet the conditions; The parameter browser provides parameter connection and visualization settings. Parameters are connected by lines to represent a specific parameter combination of a point in the trade-off space.

5. The method according to claim 1, characterized in that Step S5 is specifically as follows: different genetic algorithms are used to calculate the Pareto frontier data of target trade-off spaces of different dimensions; if the dimension of the target trade-off space is less than or equal to 3, the NSGA-II algorithm is used; if the dimension of the target trade-off space is greater than 3, the NSGA-III algorithm is used.

6. The method according to claim 1, characterized in that Step S6 specifically includes: obtaining the target touch parameters of the trade-off space display area, and adjusting the trade-off space adjustment parameters by touch; there is a mapping relationship between the touch parameters and the trade-off space adjustment parameters, and when the touch parameters change, the target trade-off space and the trade-off space adjustment parameters change synchronously.

7. The method according to claim 1, characterized in that Step S7 also includes: when a GPU is detected, redrawing is accelerated by parallel computing of the GPU.

8. The method according to claim 1, characterized in that Step S8 specifically includes: S81. Label the design solutions individually or by category, and then assign weights based on the labels; S82, manually screening out design solutions with weights greater than or equal to a specific threshold to form a solution set; S83. Further screening is performed in the solution set through set operations to determine the final trade-off solution.

9. A large-scale trade-off space parameter data processing device, characterized in that: The following units are included: A trade-off space data calculation unit, used for determining design variables of a design solution and calculating trade-off space data; A parameter storage unit, used for classifying and storing the weight space data and process parameter data; A trade-off space generating unit, used for generating an initial trade-off space according to the trade-off space data; A parameter browser generating unit, used to generate a trade-off space parameter browser according to the initial trade-off space, and the user adjusts the initial trade-off space through the trade-off space browser to obtain a target trade-off space; A Pareto front data calculation unit, used to calculate the Pareto front data of the target trade-off space; A tuning parameter adjustment unit, used to adjust the tuning parameters of the target trade-off space; A redrawing unit, used for redrawing the target trade-off space according to the adjustment parameters; The annotation unit is used to annotate the design solutions, calculate weights, and perform manual operations and set operations to obtain the final trade-off solution.

10. An electronic device, characterized in that: The method comprises a processor, a memory, a communication interface, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the processor, and the programs include instructions for executing the steps in the method according to any one of claims 1 to 8.

Images