Data processing method and apparatus

Abstract

The application discloses a data processing method and device, the method comprises the following steps: determining a specified target function corresponding to a specified target parameter in a plurality of target parameters of an initial model; determining a theoretical difference of at least one decision parameter corresponding to the specified target parameter based on the specified target function; generating a reconstructed target function corresponding to the specified target parameter based on the theoretical difference of the at least one decision parameter; and determining the value of the at least one decision parameter based on the reconstructed target function and the target functions corresponding to the target parameters other than the specified target parameter, with the condition that the specified target parameter and the target parameters other than the specified target parameter meet.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of data processing, and in particular to a data processing method and device. BACKGROUND

[0002] In a supply chain process, a multi-objective optimization model (referred to as a model) is usually used to solve the problem of material allocation to ensure reasonable allocation of materials. When using the model to allocate materials for multiple objects, due to the large amount of material data and the limitation of computer precision, the model solution result will be unreliable, so as to fail to ensure that the materials are reasonably allocated. SUMMARY

[0003] The present application provides a data processing method and device, and the present application provides the following technical solutions:

[0004] A data processing method, comprising:

[0005] determining a specified target function corresponding to a specified target parameter in a plurality of target parameters of an initial model; the initial model is based on target functions corresponding to a plurality of target parameters, and determines the value of at least one decision parameter corresponding to the plurality of target parameters, with the plurality of target parameters satisfying a condition as the goal; the target parameter is used to represent an evaluation index of material allocation; the decision parameter is used to represent an allocation attribute of material allocation;

[0006] determining a theoretical difference of at least one decision parameter corresponding to the specified target parameter based on the specified target function;

[0007] generating a reconstructed target function corresponding to the specified target parameter based on the theoretical difference of at least one decision parameter;

[0008] determining the value of at least one decision parameter based on the reconstructed target function and the target function corresponding to the target parameter other than the specified target parameter, with the specified target parameter and the target parameter other than the specified target parameter satisfying a condition as the goal.

[0009] Optionally, determining a theoretical difference of at least one decision parameter corresponding to the specified target parameter based on the specified target function comprises:

[0010] determining an optimal value of the specified target parameter based on the specified target function;

[0011] determining a theoretical difference of at least one decision parameter corresponding to the specified target parameter based on the optimal value.

[0012] Optionally, determining a theoretical difference of at least one decision parameter corresponding to the specified target parameter based on the optimal value comprises:

[0013] solving the specified target function based on the optimal value, to obtain a theoretical value of at least one decision parameter corresponding to the specified target parameter;

[0014] determining a theoretical difference of at least one decision parameter corresponding to the specified target parameter based on a gap between the actual value of at least one decision parameter corresponding to the specified target parameter and the theoretical value.

[0015] Optionally, determining a theoretical difference of at least one decision parameter corresponding to the specified target parameter based on a gap between the actual value of at least one decision parameter corresponding to the specified target parameter and the theoretical value, comprises:

[0016] determining a theoretical interval of at least one decision parameter corresponding to the specified target parameter based on the theoretical value of at least one decision parameter corresponding to the specified target parameter;

[0017] determining a theoretical difference of at least one decision parameter corresponding to the specified target parameter based on a gap between the actual value of at least one decision parameter corresponding to the specified target parameter and the theoretical interval.

[0018] Optionally, generating a reconstructed target function corresponding to the specified target parameter based on the theoretical difference of at least one decision parameter, comprises:

[0019] obtaining a theoretical difference sum based on an accumulated sum of the theoretical difference of at least one decision parameter corresponding to the specified target parameter;

[0020] generating a reconstructed target function corresponding to the specified target parameter based on the theoretical difference sum.

[0021] Optionally, generating a reconstructed target function corresponding to the specified target parameter based on the theoretical difference sum, comprises:

[0022] determining a target variable based on the theoretical difference sum; the target variable is used to represent the specified target parameter; the specified target parameter is used to represent a fairness index;

[0023] generating a reconstructed target function corresponding to the specified target parameter based on a functional relationship between the target variable and the decision parameter.

[0024] Optionally, determining a value of at least one decision parameter based on the reconstructed target function and a target function corresponding to a target parameter other than the specified target parameter, with the condition that the specified target parameter and the target parameter other than the specified target parameter satisfy a condition, comprises:

[0025] obtaining a reconstruction model; the reconstruction model is based on a target function corresponding to a specified target parameter and a target function corresponding to a target parameter other than the specified target parameter, and determines a value of at least one decision parameter, with the specified target parameter and the target parameter other than the specified target parameter satisfying a condition as an objective;

[0026] solving the reconstruction model to obtain an optimal value of at least one decision parameter.

[0027] Optionally, the specified target function corresponding to the specified target parameter in the plurality of target parameters of the initial model comprises:

[0028] According to the data magnitude of at least one target parameter in the initial model, the specified target function corresponding to the specified target parameter is obtained from the plurality of target functions.

[0029] Optionally, the specified target function corresponding to the specified target parameter in the plurality of target parameters of the initial model comprises:

[0030] According to the data magnitude of at least one target parameter in the initial model, the specified target parameter is obtained from at least one of the target parameters; the data magnitude of the specified target parameter is not within a threshold interval;

[0031] The specified target function corresponding to the specified target parameter is obtained from the target functions corresponding to the plurality of target parameters.

[0032] A data processing apparatus comprises:

[0033] A function determination unit is configured to determine a specified target function corresponding to a specified target parameter in a plurality of target parameters of an initial model; the initial model is based on target functions corresponding to the plurality of target parameters, and determines a value of at least one decision parameter corresponding to the plurality of target parameters, with the plurality of target parameters satisfying a condition as an objective; the target parameter is used to represent an evaluation index of material allocation; and the decision parameter is used to represent a distribution attribute of material allocation.

[0034] A difference determination unit is configured to determine a theoretical difference of at least one decision parameter corresponding to the specified target parameter based on the specified target function.

[0035] A function reconstruction unit is configured to generate a reconstruction target function corresponding to the specified target parameter based on the theoretical difference of at least one decision parameter.

[0036] A parameter solving unit is configured to determine a value of at least one decision parameter, with the specified target parameter and a target parameter other than the specified target parameter satisfying a condition as an objective, based on the reconstruction target function and a target function corresponding to the target parameter other than the specified target parameter.

[0037] The technical solution provided in the application determines a specified target function corresponding to a specified target parameter in a plurality of target parameters of an initial model, determines a theoretical difference of at least one decision parameter corresponding to the specified target parameter based on the specified target function, generates a reconstructed target function corresponding to the specified target parameter based on the theoretical difference of the at least one decision parameter, and determines the value of the at least one decision parameter based on the reconstructed target function and the target functions corresponding to the target parameters other than the specified target parameter, with the condition that the specified target parameter and the target parameters other than the specified target parameter meet. The application uses the theoretical difference of the at least one decision parameter corresponding to the specified target parameter to generate the reconstructed target function corresponding to the specified target parameter, so that the data magnitude of the specified target parameter is controllable, the difference in data magnitude between the specified target parameter and the target parameters other than the specified target parameter is less than a threshold, and the introduction of numerical rounding errors by the computer during operation is avoided, thereby effectively improving the reliability of the operation result and realizing the reasonable distribution of materials. BRIEF DESCRIPTION OF DRAWINGS

[0038] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, a brief introduction will be given below to the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative effort on the basis of these drawings.

[0039] Figure 1 A computer precision derivation schematic diagram provided for the embodiments of the present application;

[0040] Figure 2 A flowchart of a data processing method provided for the embodiments of the present application;

[0041] Figure 3 A flowchart of another data processing method provided for the embodiments of the present application;

[0042] Figure 4 A flowchart of still another data processing method provided for the embodiments of the present application;

[0043] Figure 5 A model reconstruction flowchart provided for the embodiments of the present application;

[0044] Figure 6 A data histogram provided for the embodiments of the present application;

[0045] Figure 7 An architecture schematic diagram of a data processing device provided for the embodiments of the present application;

[0046] Figure 8An architecture schematic diagram of an electronic device is provided in the embodiments of the present application. DETAILED DESCRIPTION

[0047] The technical solutions in the embodiments of the present application will be clearly and completely described in combination with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work fall within the scope of protection of the present application.

[0048] In the present application, the relationship terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply that there is any such actual relationship or order between these entities or operations. The term “include”, “contain” or any other variant thereof is intended to cover non-exclusive inclusion, so that the process, method, article or device including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such process, method, article or device. Without more limitations, the element defined by the statement “including a” does not exclude the presence of another identical element in the process, method, article or device including the element.

[0049] The present application scheme can be used in many general or special computing environments or configurations. For example: personal computer, server computer, handheld device or portable device, tablet device, multi-processor device, distributed computing environment including any of the above devices or devices, etc.

[0050] In the material allocation scenario shown in the background art, the decision maker of the material allocation usually pays attention to multiple evaluation indexes of the material allocation. In the case that multiple evaluation indexes meet the conditions at the same time, multiple allocation attributes of the material allocation are considered, and the allocation attribute can be specifically set as the material allocation amount of the object. For this purpose, the material allocation problem can be solved as a multi-objective optimization problem, that is, a multi-objective optimization model is used to solve the material allocation problem.

[0051] Specifically, taking the factory and product as the object of material allocation, the multiple objectives formulated by the decision maker can be: first, as many as possible to improve the product set number; second, to make the allocation of materials as fair as possible; third, the product set after allocation can maximize the profit. In the material allocation scenario, each product needs to be configured with multiple types of materials, so in addition to considering the configuration relationship between the object and the material, the fairness between the factory and the product also needs to be considered. The material demand of different factories and different products is different, and the maximum fairness is taken as the target. Usually, the difference between the satisfaction rate of each object (i.e. the ratio of the material allocation amount to the material demand amount) is taken as the measurement standard, that is, the smaller the difference, the higher the fairness index.

[0052] Under the condition of meeting the configuration relationship between the object and the material, the sum of the material allocation amount of each object needs to be equal to the total supply amount of the material, and the data magnitude difference between the total supply amount of the material and the material demand of each object is less than a threshold value, for example, the total supply amount of the material is 2×10 9 kg, and the minimum material demand of the object is 3×10 1 kg, the data magnitude difference between the two is 10 8 , and based on the computer precision, the threshold value is 10 12 , and the data magnitude difference between the two is less than the threshold value (the threshold value is used to represent the computer precision).

[0053] However, the applicant found that when using the model to allocate materials for multiple objects, the data magnitude difference between multiple target parameters in the model solving process is greater than the threshold value, thereby introducing numerical rounding errors during operation, resulting in unreliable model solving results.

[0054] Specifically, when optimizing the maximum fairness, the calculation of the satisfaction rate is involved. Based on the definition of the satisfaction rate, the satisfaction rate involves division. In the case of a large data magnitude difference between the material allocation amount and the material demand amount, the data magnitude of the satisfaction rate will be greater than the threshold value, thereby causing the data magnitude difference between the fairness index and other evaluation indexes to be greater than the threshold value. For example, the data magnitude of the material allocation is 10 -2 , the data magnitude of the material demand is 10 6 , and the corresponding data magnitude of the satisfaction rate is 10 -8 , that is, the data magnitude of the fairness index is 10 -8 , and the data magnitude of the profit index is 10 8 , and the data magnitude difference between the fairness index and the profit index is 10 16 , which is much larger than the threshold value defined by the computer precision, thereby introducing numerical rounding errors during the solving of the multi-objective optimization model, resulting in unreliable solving results (i.e. the material allocation amount of each object) of the multi-objective optimization model.

[0055] It should be noted that the computer itself is represented by a limited number of bit patterns to represent an infinite number of real numbers, and numerical rounding errors will always be introduced, so the computer precision is set, and when the data magnitude difference involved in the operation is greater than the computer precision, the result obtained by the operation will introduce numerical rounding errors, which will lead to the inaccuracy of the operation result. The derivation process of the operation result caused by the computer precision problem can be simply summarized as Figure 1

[0056] To sum up, in order to avoid introducing numerical rounding errors when solving the multi-objective optimization model, resulting in unreliable model solving results, the present application provides a data processing method for controlling the data magnitude of special evaluation indicators (i.e. evaluation indicators whose data magnitude changes greatly due to algorithm, such as fairness indicators) in material allocation, ensuring that the data magnitude difference between special evaluation indicators and other evaluation indicators is less than the computer precision.

[0057] It should be noted that the data processing method provided by the embodiments of the present application can be applied to various system platforms, and the execution subject thereof includes but is not limited to computers, terminals and processors.

[0058] Optionally, as shown in Figure 2 , a flowchart of a data processing method provided by the embodiments of the present application is shown, which includes the following steps.

[0059] S201: Determine the specified target function corresponding to the specified target parameter in the plurality of target parameters of the initial model.

[0060] The initial model is based on the target functions corresponding to the plurality of target parameters, and determines the value of at least one decision parameter corresponding to the plurality of target parameters, with the plurality of target parameters satisfying the condition as the goal. The target parameter is used to represent the evaluation indicator of material allocation, and the decision parameter is used to represent the allocation attribute of material allocation.

[0061] The so-called initial model can be understood as a multi-objective optimization model for solving material allocation. In the embodiments of the present application, the specified target parameter can be used to represent the special evaluation indicator mentioned above, such as the fairness indicator. In order to ensure that the data magnitude difference between the specified target parameter and the target parameter other than the specified target parameter is less than the threshold value defined by the computer precision, it is necessary to reconstruct the specified target function corresponding to the specified target parameter to ensure that the data magnitude of the specified target parameter is controllable. In addition, the condition satisfied by the plurality of target parameters can be understood as the condition that the plurality of evaluation indicators of material allocation need to reach, such as the condition of ensuring that the fairness indicator is maximized, the profit indicator is maximized, etc.

[0062] ​It should be noted that before reconstructing the specified target function corresponding to the specified target parameter, the specified target parameter needs to be selected from the plurality of target parameters of the initial model. In the embodiments of the present application, the specified target function corresponding to the specified target parameter can be obtained from the plurality of target functions according to the data magnitude of at least one target parameter in the initial model.

[0063] Optionally, the specific implementation process of obtaining the specified target function corresponding to the specified target parameter from the plurality of target functions according to the data magnitude of at least one target parameter in the initial model can include: obtaining the specified target parameter from at least one target parameter according to the data magnitude of at least one target parameter in the initial model; the data magnitude of the specified target parameter is not within the threshold interval; and obtaining the specified target function corresponding to the specified target parameter from the target functions corresponding to the plurality of target parameters.

[0064] Specifically, the plurality of target parameters involved in the initial model include a first target parameter, a second target parameter and a third target parameter, the first target parameter is used to represent a fairness index of material allocation, the second target parameter is used to represent a profit index of material allocation, and the third target parameter is used to represent a completeness index of material allocation. In order to ensure that the data magnitude difference between the fairness index, the profit index and the completeness index is less than a threshold value, the data magnitudes of the first target parameter, the second target parameter and the third target parameter are obtained in advance. Assuming that the data magnitude of the first target parameter is not within the threshold interval, it means that the data magnitude difference between the first target parameter and the second target parameter and the data magnitude difference between the first target parameter and the third target parameter will be greater than the threshold value. Therefore, the first target parameter is identified as the specified target parameter, and the target function corresponding to the first target parameter is identified as the specified target function.

[0065] S202: Determine the theoretical difference of at least one decision parameter corresponding to the specified target parameter based on the specified target function.

[0066] In the field of target optimization, the essence of solving the target function is to determine the value of at least one decision parameter corresponding to the target parameter based on the function extreme value (i.e. the optimal value of the target parameter) of the target function. Obviously, in the case that the specified target parameter reaches the optimal value, the decision parameter corresponding to the specified target parameter will also reach the optimal value. Therefore, the theoretical difference of the decision parameter corresponding to the specified target parameter can be used instead of the function extreme value of the specified target function as the optimization target.

[0067] It should be noted that the specific implementation process of determining the theoretical difference of at least one decision parameter corresponding to the specified target parameter based on the specified target function can refer to the steps shown in Figure 3 and Figure 4 .

[0068] S203: generating a reconstruction objective function corresponding to the specified target parameter based on the theoretical difference of the at least one decision parameter.

[0069] In order to ensure that the data magnitude difference between the specified target parameter and the target parameter other than the specified target parameter is less than the threshold value, it is necessary to ensure that the data magnitude of the specified target parameter is kept within the threshold interval. Generally, the main factor causing the data magnitude of the specified target parameter not to be within the threshold interval is usually caused by the algorithm involved in the specified target function, for example, the specified target function corresponding to the fairness index involves division of a large data magnitude, resulting in the data magnitude of the specified target parameter corresponding to the fairness index being less than the upper limit value of the threshold interval.

[0070] In the embodiments of the present application, the specified target function can be reconstructed to ensure that the data magnitude of the target parameter is within the threshold interval. Specifically, the reconstruction target function can be generated based on the theoretical difference of the at least one decision parameter corresponding to the specified target parameter to realize the reconstruction of the specified target function. Since the calculation process of the theoretical difference does not involve algorithms with large data magnitudes (such as multiplication, division, and square root), the data magnitude of the specified target parameter can be ensured not to change greatly, and can be consistent with the data magnitude of the decision parameter corresponding to the specified target parameter. In addition, the data magnitude of the decision parameter is by default within the threshold interval, so the data magnitude of the specified target parameter can be ensured to be within the threshold interval.

[0071] It should be noted that the specific implementation process of generating a reconstruction objective function corresponding to the specified target parameter based on the theoretical difference of the at least one decision parameter can refer to the steps shown in Figure 3 and Figure 4 .

[0072] S204: determining the value of the at least one decision parameter based on the reconstruction objective function and the objective function corresponding to the target parameter other than the specified target parameter, so that the specified target parameter and the target parameter other than the specified target parameter satisfy the condition.

[0073] The reconstruction objective function and the objective function corresponding to the target parameter other than the specified target parameter can form a reconstruction model, and the multi-objective optimization function of the reconstruction model and the initial model is consistent, and they are both used for material allocation of each object. The difference is that the numerical rounding error is not introduced when solving the reconstruction model, that is, the reliability of the solution result of the reconstruction model is higher than that of the initial model, and the reasonable allocation of materials can be ensured.

[0074] To this end, the specific implementation process of determining the value of the at least one decision parameter based on the reconstructed objective function and the objective function corresponding to the target parameter other than the specified target parameter, with the condition that the specified target parameter and the target parameter other than the specified target parameter are satisfied, as the target, can be: obtaining a reconstruction model; the reconstruction model determines the value of the at least one decision parameter based on the reconstructed objective function and the objective function corresponding to the target parameter other than the specified target parameter, with the condition that the specified target parameter and the target parameter other than the specified target parameter are satisfied, as the target; solving the reconstruction model to obtain the optimal value of the at least one decision parameter.

[0075] It should be noted that the optimal value of the at least one decision parameter represents the optimal value of each allocation attribute of the material allocation, and in the material allocation scenario shown in the embodiments of the present application, it can also be understood as: under the condition that the plurality of evaluation indexes of the material allocation satisfy the condition, the material allocation amount of each object.

[0076] Based on the processes S201-S204, the embodiments of the present application can generate a reconstructed objective function corresponding to the specified target parameter using the theoretical difference of the at least one decision parameter corresponding to the specified target parameter, so that the data magnitude of the specified target parameter is controllable, and the data magnitude difference between the specified target parameter and the target parameter other than the specified target parameter is less than a threshold value, thereby avoiding the introduction of numerical rounding errors during computer operation, effectively improving the reliability of the operation result, and achieving reasonable allocation of materials.

[0077] Optionally, as shown in Figure 3 Another flowchart of a data processing method provided by the embodiments of the present application is shown, which includes the following steps.

[0078] S301: determining the optimal value of the specified target parameter based on the specified objective function.

[0079] In the material allocation scenario mentioned in the embodiments of the present application, the constant parameters involved in the specified objective function (for example, the fairness index, the constant parameters include the total supply of materials, the number of material allocation objects, and the material demand of each object) are known, therefore, under the condition that the constant parameters and the condition satisfied by the specified target parameter are known, the optimal value of the specified target parameter can be determined based on the specified objective function.

[0080] S302: based on the optimal value, solving the specified objective function to obtain the theoretical value of the at least one decision parameter corresponding to the specified target parameter.

[0081] It can be understood that, in the case that the constant parameters involved in the specified target function and the conditions met by the specified target parameters are known, the optimal value of the specified target parameter is substituted into the specified target function, and the specified target function is solved, and the value of the decision parameter corresponding to the specified target parameter obtained is the theoretical value of the decision parameter.

[0082] S303: Based on the gap between the actual value and the theoretical value of the at least one decision parameter corresponding to the specified target parameter, the theoretical difference of the at least one decision parameter corresponding to the specified target parameter is determined.

[0083] Wherein, the gap between the actual value and the theoretical value of the decision parameter can be specifically represented as: calculating the absolute value of the difference between the actual value and the theoretical value of the decision parameter.

[0084] S304: Based on the cumulative sum of the theoretical difference of the at least one decision parameter corresponding to the specified target parameter, the theoretical difference sum is obtained.

[0085] Wherein, the theoretical difference sum can be regarded as the cumulative sum of the theoretical difference of the plurality of decision parameters, and generally, the theoretical difference sum is consistent with the data magnitude of the decision parameter.

[0086] Therefore, after obtaining the theoretical difference sum, the reconstruction target function corresponding to the specified target parameter can be generated based on the theoretical difference sum, so as to ensure that the data magnitude of the specified target parameter is within the threshold interval.

[0087] Further, the specific implementation process of generating the reconstruction target function corresponding to the specified target parameter based on the theoretical difference sum can be referred to S305-S306.

[0088] S305: Based on the theoretical difference sum, the target variable is determined.

[0089] Wherein, the target variable is used to represent the specified target parameter, and in the embodiments of the present application, the specified target parameter can be specifically used to represent the fairness index.

[0090] It should be noted that the fairness index represented by the specified target parameter is only a specific form of special evaluation index, and the evaluation index whose data magnitude is not within the threshold interval can be regarded as the specified target parameter.

[0091] S306: Based on the functional relationship between the target variable and the decision parameter, the reconstruction target function corresponding to the specified target parameter is generated.

[0092] It can be understood that the functional relationship between the target variable and the decision parameter belongs to the common sense of the mathematical field, which will not be repeated here.

[0093] To facilitate understanding of the specific generation process of the reconstructed objective function shown in the embodiments of the present application, the embodiments of the present application take the fairness index as an example to expand and explain, and the specific explanation content is as follows.

[0094] Suppose that the specified target function corresponding to the specified target parameter representing the fairness index is as shown in formula (1).

[0095]

[0096] In formula (1), minimize represents a function of finding the minimum value, s.t. represents a constraint condition, x i represents the decision parameter (specifically, the material allocation amount of each factory) corresponding to the specified target parameter, i represents the index of the decision parameter (which can represent the index of the factory), n represents the total number of decision parameters, s represents a constant parameter (specifically, the total supply amount of materials), d i represents a constant parameter (specifically, the material demand amount of each factory), represents the satisfaction rate of each factory, represents the maximum satisfaction rate (i.e., the satisfaction rate with the maximum value among the satisfaction rates), represents the minimum satisfaction rate (i.e., the satisfaction rate with the minimum value among the satisfaction rates).

[0097] As can be seen from formula (1), the smaller the gap between the maximum satisfaction rate and the minimum satisfaction rate, the greater the fairness index, and to pursue the maximum fairness as the target, the gap between the maximum satisfaction rate and the minimum satisfaction rate should theoretically be 0.

[0098] Firstly, based on the specified target function shown in formula (1), the optimal value of the specified target parameter can be determined, and based on the optimal value, formula (1) is solved to obtain the theoretical value i of at least one decision parameter x

[0099] Secondly, based on the gap between the actual value and the theoretical value of at least one decision parameter corresponding to the specified target parameter, the theoretical difference of at least one decision parameter corresponding to the specified target parameter is determined.

[0100] Then, based on the cumulative sum of the theoretical difference of at least one decision parameter corresponding to the specified target parameter, the theoretical difference total sum

[0101] Finally, based on the theoretical difference total sum as the target variable, and based on the functional relationship between the target variable and the decision parameter, a reconstructed objective function is generated, and the specific expression of the reconstructed objective function can be seen from formula (2).

[0102]

[0103] Based on formula (2), the algorithm involving large data magnitude in the reconstruction objective function (i.e., the calculation process of the satisfaction rate) is no longer involved, so that the data magnitude of the specified target parameter and the decision parameter can be ensured to be consistent, so that the data magnitude of the specified target parameter is within the threshold interval.

[0104] It should be noted that for the specified target function involving the algorithm of large data magnitude, the theoretical difference of at least one decision parameter corresponding to the specified target parameter can be used to generate a corresponding reconstruction objective function to replace the specified target function.

[0105] Based on the above-mentioned processes S301-S306, the reconstruction objective function corresponding to the specified target parameter can be generated based on the theoretical difference of at least one decision parameter corresponding to the specified target parameter, so that the data magnitude of the specified target parameter is controllable.

[0106] Optionally, as shown in Figure 4 The flowchart of another data processing method provided by the embodiment of the present application is shown in the figure.

[0107] S401: determining the optimal value of the specified target parameter based on the specified target function.

[0108] The specific implementation principle of S401 can be referred to the step explanation and description of S301 described above, which will not be repeated here.

[0109] S402: solving the specified target function based on the optimal value to obtain the theoretical value of at least one decision parameter corresponding to the specified target parameter.

[0110] The specific implementation principle of S402 can be referred to the step explanation and description of S302 described above, which will not be repeated here.

[0111] S403: determining the theoretical interval of at least one decision parameter corresponding to the specified target parameter based on the theoretical value of at least one decision parameter corresponding to the specified target parameter.

[0112] The theoretical value of the decision parameter obtained based on the step S402 may have a slight deviation, in order to avoid the deviation of the theoretical value, the theoretical interval of the decision parameter is determined based on the theoretical value.

[0113] Specifically, the upper limit value of the theoretical value can be determined based on the upward floating ratio of the theoretical value, the lower limit value of the theoretical value can be determined based on the downward floating ratio of the theoretical value, and the value interval between the upper limit value and the lower limit value is identified as the theoretical interval of the decision parameter. The specific values of the upward floating ratio and the downward floating ratio can be set by the technician according to the actual situation.

[0114] S404: determining a theoretical difference of the at least one decision parameter corresponding to the specified target parameter based on a gap between the actual value of the at least one decision parameter corresponding to the specified target parameter and the theoretical interval.

[0115] It can be understood that, compared with the gap between the actual value and the theoretical value of the decision parameter mentioned in S303, as the theoretical difference of the decision parameter, the actual value of the decision parameter and the gap between the theoretical interval, as the theoretical difference of the decision parameter, based on the embodiment of the application, has higher reliability.

[0116] It should be noted that the gap between the actual value and the theoretical interval of the decision parameter can be specifically represented as: the maximum value of the first value, the second value and the limited value is identified as the gap between the actual value and the theoretical interval, the first value includes the difference between the actual value and the upper limit value of the theoretical interval, the second value includes the difference between the actual value and the lower limit value of the theoretical interval, and the limited value is set to 0.

[0117] S405: obtaining a theoretical difference sum based on the cumulative sum of the theoretical difference of the at least one decision parameter corresponding to the specified target parameter.

[0118] The specific implementation principle of S405 can be referred to the step explanation of S304 described above, which will not be repeated here.

[0119] S406: generating a reconstructed target function corresponding to the specified target parameter based on the theoretical difference sum.

[0120] The specific implementation principle of S406 can be referred to the step explanation of S305-S306 described above, which will not be repeated here.

[0121] In order to facilitate understanding of the specific generation process of the reconstructed target function shown in the embodiment of the application, the embodiment of the application takes the specified target function shown in formula (1) as an example to explain this, and the specific explanation content is as follows.

[0122] Firstly, based on the specified target function shown in formula (1), the optimal value of the specified target parameter can be determined, and based on the optimal value, formula (1) is solved to obtain the theoretical value of the at least one decision parameter x i corresponding to the specified target parameter.

[0123] Secondly, based on the theoretical value of the at least one decision parameter corresponding to the specified target parameter, the theoretical interval of the at least one decision parameter corresponding to the specified target parameter is determined In the theoretical interval, represents the upper limit value of the theoretical interval, a lower limit value representing a theoretical interval.

[0124] Then, based on the gap between the actual value of the at least one decision parameter corresponding to the specified target parameter and the theoretical interval, a theoretical difference of the at least one decision parameter corresponding to the specified target parameter is determined

[0125]

[0126] Subsequently, based on the cumulative sum of the theoretical differences of the at least one decision parameter corresponding to the specified target parameter, a theoretical difference sum is obtained

[0127] Finally, based on the theoretical difference sum, a reconstructed target function corresponding to the specified target parameter is generated, and the specific expression of the reconstructed target function can be seen from formula (3).

[0128]

[0129] In formula (3), max represents a maximum value function, p represents a floating ratio, specifically, (1+p) represents an upward floating ratio of the theoretical value, and (1-p) represents a downward floating ratio of the theoretical value.

[0130] Based on formula (3), it can be seen that the reconstructed target function no longer involves an algorithm with a large data order of magnitude (i.e., the calculation process of the satisfaction rate), and therefore, the data order of magnitude of the specified target parameter and the decision parameter can be ensured to be consistent, so that the data order of magnitude of the specified target parameter is within the threshold interval.

[0131] Based on the above-mentioned processes S401-S406, the embodiment of the present application can generate a reconstructed target function corresponding to the specified target parameter based on the theoretical difference of the at least one decision parameter corresponding to the specified target parameter, so that the data order of magnitude of the specified target parameter is controllable.

[0132] It should be noted that, for the reconstructed target functions shown in formula (2) and formula (3), the constraint condition shown in formula (2) can be regarded as a point constraint, and the constraint condition shown in formula (3) can be regarded as an interval constraint. The constraints brought by the point constraint and the interval constraint for the reconstructed target function are both linear constraints, which can reduce the overall constraint amount of the reconstructed target function and effectively improve the solving efficiency of the reconstructed target function.

[0133] Therefore, in order to understand the difference between the point constraint and the interval constraint of the reconstructed target function, the embodiment of the present application is based on the model reconstruction flowchart shown in formula (4) for a simple description, and the specific implementation process is as follows. Figure 5

[0134] 1. Based on the initial model of the material distribution scene, a specified target function corresponding to the specified target parameter is obtained.​

[0135] 2. determining a theoretical value of the at least one decision parameter corresponding to the specified target parameter based on the specified target function.

[0136] 3. determining the point constraint and the interval constraint based on the theoretical value of the at least one decision parameter.

[0137] 4. reconstructing the specified target function by using the point constraint to obtain a first reconstructed target function.

[0138] 5. adding the interval constraint in the first reconstructed target function to obtain a second reconstructed target function.

[0139] 6. using the second reconstructed target function to replace the specified target function in the multi-objective optimization model to obtain a reconstructed model.

[0140] 7. solving the reconstructed model to obtain an optimal value of the at least one decision parameter.

[0141] Specifically, taking the fairness index as an example, the difference between the point constraint and the interval constraint can also be seen from the histogram shown in Figure 6 Generally, compared with the point constraint, the interval constraint can further improve the solving quality of the reconstructed target function, so that the solving result of the reconstructed target function is more reasonable.

[0142] In summary, the reconstructed target function shown in the embodiments of the present application has higher reliability of the solving result compared with the specified target function, and can ensure reasonable allocation of materials.

[0143] Corresponding to the data processing method provided by the above embodiments of the present application, the embodiments of the present application also provide a data processing device.

[0144] Optionally, as shown in Figure 7 , an architecture schematic diagram of a data processing device provided by the embodiments of the present application includes the following units.

[0145] The function determination unit 100 is configured to determine a specified target function corresponding to a specified target parameter in a plurality of target parameters of an initial model; the initial model is based on a plurality of target functions corresponding to the plurality of target parameters, and determines a value of at least one decision parameter corresponding to the plurality of target parameters, with the plurality of target parameters satisfying a condition as a target; the target parameter is used to represent an evaluation index of material allocation; and the decision parameter is used to represent an allocation attribute of material allocation.

[0146] In the data processing device shown in the embodiments of the present application, the function determination unit 100 can be specifically configured to: according to a data magnitude of at least one target parameter in the initial model, acquire the specified target function corresponding to the specified target parameter from the plurality of target functions.

[0147] Optionally, the function determining unit 100 is specifically configured to: obtain a specified target parameter from the at least one target parameter according to a data magnitude of the at least one target parameter in the initial model; the data magnitude of the specified target parameter is not within a threshold interval; and obtain a specified target function corresponding to the specified target parameter from the target functions corresponding to the plurality of target parameters.

[0148] The difference determining unit 200 is configured to determine a theoretical difference of at least one decision parameter corresponding to the specified target parameter based on the specified target function.

[0149] In the data processing apparatus shown in the embodiments of the present application, the difference determining unit 200 is specifically configured to: determine an optimal value of the specified target parameter based on the specified target function; and determine a theoretical difference of at least one decision parameter corresponding to the specified target parameter based on the optimal value.

[0150] Optionally, the difference determining unit 200 is specifically configured to: obtain a theoretical value of at least one decision parameter corresponding to the specified target parameter by solving the specified target function based on the optimal value; and determine a theoretical difference of at least one decision parameter corresponding to the specified target parameter based on a difference between an actual value of at least one decision parameter corresponding to the specified target parameter and the theoretical value.

[0151] Optionally, the difference determining unit 200 is specifically configured to: determine a theoretical interval of at least one decision parameter corresponding to the specified target parameter based on the theoretical value of at least one decision parameter corresponding to the specified target parameter; and determine a theoretical difference of at least one decision parameter corresponding to the specified target parameter based on a difference between an actual value of at least one decision parameter corresponding to the specified target parameter and the theoretical interval.

[0152] The function reconstructing unit 300 is configured to generate a reconstructed target function corresponding to the specified target parameter based on the theoretical difference of at least one decision parameter.

[0153] In the data processing apparatus shown in the embodiments of the present application, the function reconstructing unit 300 is specifically configured to: obtain a theoretical difference total sum based on an accumulated sum of the theoretical difference of at least one decision parameter corresponding to the specified target parameter; and generate a reconstructed target function corresponding to the specified target parameter based on the theoretical difference total sum.

[0154] Optionally, the function reconstructing unit 300 is specifically configured to: determine a target variable based on the theoretical difference total sum; the target variable is used to represent the specified target parameter; the specified target parameter is used to represent a fairness index; and generate a reconstructed target function corresponding to the specified target parameter based on a functional relationship between the target variable and the decision parameter.

[0155] The parameter solving unit 400 is configured to determine the value of the at least one decision parameter based on the reconstructed objective function and the objective function corresponding to the target parameter other than the specified target parameter, so that the specified target parameter and the target parameter other than the specified target parameter satisfy the condition.

[0156] In the data processing apparatus, the parameter solving unit 400 can be specifically configured to: obtain a reconstruction model; determine the value of the at least one decision parameter based on the reconstructed objective function and the objective function corresponding to the target parameter other than the specified target parameter, so that the specified target parameter and the target parameter other than the specified target parameter satisfy the condition; and solve the reconstruction model to obtain the optimal value of the at least one decision parameter.

[0157] Based on the units of the data processing apparatus, the embodiments of the present application can generate the reconstructed objective function corresponding to the specified target parameter by using the theoretical difference of the at least one decision parameter corresponding to the specified target parameter, so that the data magnitude of the specified target parameter is controllable, and the difference between the data magnitudes of the specified target parameter and the target parameter other than the specified target parameter is less than the threshold value, thereby avoiding the introduction of numerical rounding errors in the operation of the computer, effectively improving the reliability of the operation result, and realizing the reasonable allocation of materials.

[0158] The present application also provides a computer-readable storage medium, which includes a stored program, wherein the program executes the data processing method provided by the embodiments of the present application.

[0159] The present application also provides an electronic device, such as Figure 8 As shown, the electronic device includes a processor, a memory and a bus. The processor is connected with the memory through the bus. The memory is configured to store a program, and the processor is configured to run the program. When the program is running, the data processing method provided by the embodiments of the present application is executed.

[0160] In addition, the functions described above in the embodiments of the present application can be at least partially performed by one or more hardware logic components. For example, non-limiting examples of exemplary types of hardware logic components that can be used include: field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), system on a chip (SOCs), complex programmable logic devices (CPLDs), etc.

[0161] Although the subject matter has been described in language specific to structural features and / or methodological acts, it is to be understood that the subject defined in the appended claims is not necessarily limited to the specific features or acts described above. Rather, the specific features and acts described above are disclosed as example forms of implementing the claims.

[0162] While several inventive embodiments have been described above, it should be understood that many variations and modifications of the inventive concepts described above can be made without departing from the scope of the application. Accordingly, the domain of the application is not limited to the specific embodiments described above, but only by the claims that follow.

[0163] The above description is merely illustrative of the application and the inventive concepts presented within this specification. Depending on the embodiment, certain features of the application can be employed without the requirement of other features. Accordingly, the domain of the application is not limited to only those embodiments described above, but rather includes all alternatives, modifications, and equivalents falling within the scope of the claims along with their full scope of equivalents.

Claims

1. A data processing method, comprising: determining a specified objective function corresponding to a specified target parameter in a plurality of target parameters of an initial model; determining a value of at least one decision parameter corresponding to the plurality of target parameters of the initial model based on objective functions corresponding to the plurality of target parameters, with the plurality of target parameters satisfying a condition as a target, the target parameters being used to represent evaluation indexes of material allocation, the decision parameters being used to represent allocation attributes of material allocation, a data magnitude of the specified target parameter not being in a threshold interval; determining a theoretical difference of the at least one decision parameter corresponding to the specified target parameter based on the specified objective function; generating a reconstructed objective function corresponding to the specified target parameter based on the theoretical difference of the at least one decision parameter; determining a value of the at least one decision parameter based on the reconstructed objective function and objective functions corresponding to target parameters other than the specified target parameter, with the specified target parameter and the target parameters other than the specified target parameter satisfying a condition as a target.

2. The method of claim 1, wherein determining the theoretical difference of the at least one decision parameter corresponding to the specified target parameter based on the specified objective function comprises: determining an optimal value of the specified target parameter based on the specified objective function; determining the theoretical difference of the at least one decision parameter corresponding to the specified target parameter based on the optimal value.

3. The method of claim 2, wherein determining the theoretical difference of the at least one decision parameter corresponding to the specified target parameter based on the optimal value comprises: solving the specified objective function based on the optimal value to obtain a theoretical value of the at least one decision parameter corresponding to the specified target parameter; determining the theoretical difference of the at least one decision parameter corresponding to the specified target parameter based on a gap between an actual value of the at least one decision parameter corresponding to the specified target parameter and the theoretical value.

4. The method of claim 3, wherein determining the theoretical difference of the at least one decision parameter corresponding to the specified target parameter based on the gap between the actual value of the at least one decision parameter corresponding to the specified target parameter and the theoretical value comprises: determining a theoretical interval of the at least one decision parameter corresponding to the specified target parameter based on the theoretical value of the at least one decision parameter corresponding to the specified target parameter; determining the theoretical difference of the at least one decision parameter corresponding to the specified target parameter based on a gap between the actual value of the at least one decision parameter corresponding to the specified target parameter and the theoretical interval.

5. The method of claim 1, wherein generating the reconstructed objective function corresponding to the specified target parameter based on the theoretical difference of the at least one decision parameter comprises: obtaining a theoretical difference sum based on an accumulated sum of the theoretical difference of the at least one decision parameter corresponding to the specified target parameter; generating the reconstructed objective function corresponding to the specified target parameter based on the theoretical difference sum.

6. The method of claim 5, wherein generating the reconstructed objective function corresponding to the specified target parameter based on the theoretical difference sum comprises: determine a target variable based on the total difference; the target variable is used to represent the specified target parameter; the specified target parameter is used to represent the fairness index; generate a reconstructed target function corresponding to the specified target parameter based on the functional relationship between the target variable and the decision parameter.

7. The method of claim 1, wherein the value of the at least one decision parameter is determined based on the reconstructed target function and the target function corresponding to the target parameter other than the specified target parameter, with the specified target parameter and the target parameter other than the specified target parameter satisfying the condition as the goal, comprising: obtaining a reconstruction model; the reconstruction model is based on the reconstructed target function and the target function corresponding to the target parameter other than the specified target parameter, with the specified target parameter and the target parameter other than the specified target parameter satisfying the condition as the goal, to determine the value of the at least one decision parameter; solving the reconstruction model to obtain the optimal value of the at least one decision parameter.

8. The method of claim 1, wherein the specified target function corresponding to the specified target parameter of the initial model is determined, comprising: According to the data magnitude of at least one target parameter in the initial model, the specified target function corresponding to the specified target parameter is obtained from a plurality of target functions.

9. The method of claim 8, wherein the specified target function corresponding to the specified target parameter is obtained from a plurality of target functions according to the data magnitude of at least one target parameter in the initial model, comprising: According to the data magnitude of at least one target parameter in the initial model, the specified target parameter is obtained from at least one of the target parameters; the specified target function corresponding to the specified target parameter is obtained from a plurality of target functions corresponding to the target parameters.

10. A data processing apparatus, comprising: a function determination unit configured to determine a specified target function corresponding to a specified target parameter of a plurality of target parameters of an initial model; the initial model is based on a plurality of target functions corresponding to a plurality of target parameters, with the plurality of target parameters satisfying the condition as the goal, to determine the value of at least one decision parameter corresponding to the plurality of target parameters; the target parameter is used to represent the evaluation index of material allocation; the decision parameter is used to represent the allocation attribute of material allocation; the data magnitude of the specified target parameter is not within the threshold interval; a difference determination unit configured to determine a theoretical difference of at least one decision parameter corresponding to the specified target parameter based on the specified target function; a function reconstruction unit configured to generate a reconstructed target function corresponding to the specified target parameter based on the theoretical difference of at least one decision parameter; a parameter solving unit configured to determine the value of at least one decision parameter based on the reconstructed target function and the target function corresponding to the target parameter other than the specified target parameter, with the specified target parameter and the target parameter other than the specified target parameter satisfying the condition as the goal.

Images