Data processing method and device, electronic equipment and computer readable storage medium

Abstract

The invention provides a data processing method and device, and relates to the technical field of data processing, in particular to the technical fields of machine learning, large language models and the like. The specific implementation scheme is as follows: obtaining object attribute data of a plurality of objects in a target activity scene, wherein the object attribute data is used for representing object features of each object in the target activity scene; obtaining an operator body and configuration parameters of an algorithm operator; determining a genetic algorithm based on the operator ontology and the configuration parameters; based on the object attribute data, performing group division on the plurality of objects by adopting a genetic algorithm to obtain a group division result; and based on a group division result, sending grouping information to the plurality of objects, so that the plurality of objects move in the target activity scene based on the grouping information.

Description

Technical Field

[0001] The present disclosure relates to the field of data processing technology, particularly to machine learning and large-scale language models, and more particularly to a data processing method and apparatus, an electronic device, and a computer-readable storage medium. Background Art

[0002] Genetic algorithms are a classic approach to optimization and search problems, based on the principles of natural selection and genetics. They gradually search for optimal solutions through operations such as selection, mating, and mutation.

[0003] In terms of group formation, traditional genetic algorithms have been used for team building and group optimization. However, in certain specific scenarios, the adaptability and performance of genetic algorithms are limited. They are prone to falling into local optimal solutions and lack adaptability to different input data. In addition, existing genetic algorithms often face the problem of inefficient group division. Summary of the Invention

[0004] The present disclosure provides a data processing method and apparatus, an electronic device, and a computer-readable storage medium.

[0005] According to a first aspect, a data processing method is provided, which includes: obtaining object attribute data of multiple objects in a target activity scene, the object attribute data being used to characterize the object characteristics of each object in the target activity scene; obtaining an operator ontology and configuration parameters of a user algorithm operator; determining a genetic algorithm based on the operator ontology and the configuration parameters; based on the object attribute data, using a genetic algorithm to divide the multiple objects into groups to obtain group division results; based on the group division results, sending grouping information to the multiple objects, so that the multiple objects are active in the target activity scene based on the grouping information.

[0006] According to a second aspect, a data processing device is provided, which includes: an attribute acquisition unit, configured to acquire object attribute data of multiple objects in a target activity scene, the object attribute data being used to characterize object characteristics of each object in the target activity scene; a parameter acquisition unit, configured to acquire an operator ontology and configuration parameters of an algorithm operator; a determination unit, configured to determine a genetic algorithm based on the operator ontology and the configuration parameters; a division unit, configured to use a genetic algorithm to perform group division on multiple objects based on the object attribute data to obtain a group division result; and a grouping unit, configured to send grouping information to multiple objects based on the group division result, so that the multiple objects are active in the target activity scene based on the grouping information.

[0007] According to a third aspect, an electronic device is provided, comprising: at least one processor; and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the method described in any implementation manner of the first aspect.

[0008] According to a fourth aspect, a non-transitory computer-readable storage medium storing computer instructions is provided, where the computer instructions are used to cause a computer to execute the method as described in any implementation of the first aspect.

[0009] The data processing method and apparatus provided by the embodiments of the present disclosure first obtain object attribute data of multiple objects in a target activity scene, where the object attribute data is used to characterize the object characteristics of each object in the target activity scene; secondly, obtain the operator ontology and configuration parameters of the algorithm operator; thirdly, determine a genetic algorithm based on the operator ontology and configuration parameters; thirdly, use the genetic algorithm to group the multiple objects based on the object attribute data to obtain a group division result; and finally, based on the group division result, send grouping information to the multiple objects so that the multiple objects can be active in the target activity scene based on the grouping information. Thus, through the user's parameter setting of the algorithm operator of the genetic algorithm, the genetic algorithm can flexibly adapt to different types of object attribute data, improve the genetic algorithm's ability to perform comprehensive analysis of multiple factor data, improve the generalization ability and adaptability of the genetic algorithm, and improve the efficiency of group division.

[0010] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present disclosure, nor are they intended to limit the scope of the present disclosure. Other features of the present disclosure will become readily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] The accompanying drawings are provided to facilitate a better understanding of the present invention and do not constitute a limitation of the present disclosure.

[0012] Figure 1 is a flow chart of an embodiment of a data processing method according to the present disclosure;

[0013] Figure 2 is a schematic diagram of the data structure of a chromosome in the present disclosure;

[0014] Figure 3 is a schematic structural diagram of an embodiment of a data processing device according to the present disclosure;

[0015] Figure 4 It is a block diagram of an electronic device used to implement the data processing method according to an embodiment of the present disclosure. DETAILED DESCRIPTION

[0016] Unless expressly stated otherwise, throughout the specification and claims, the term "comprise" or variations such as "include" or "comprising", etc., will be understood to include the stated elements or components but not to exclude other elements or other components.

[0017] The technical solutions of the present disclosure are described below through specific examples. It should be understood that one or more steps mentioned in the present disclosure do not exclude the existence of other methods and steps before and after the combination step, or other methods and steps may be inserted between these explicitly mentioned steps. It should also be understood that these examples are only used to illustrate the present disclosure and are not used to limit the scope of the present disclosure. Unless otherwise specified, the numbering of each method step is only for the purpose of identifying each method step, and does not limit the order of arrangement of each method or limit the scope of implementation of the present disclosure. Changes or adjustments in their relative relationships can also be regarded as the scope of implementation of the present disclosure without substantial changes in the technical content.

[0018] The sources of the raw materials and instruments used in the examples are not particularly limited and can be purchased from the market or prepared according to conventional methods known to those skilled in the art.

[0019] In response to the defects of group division in traditional technologies, the present disclosure proposes a data processing method, which determines a genetic algorithm based on configuration parameters configured by the user, and uses the genetic algorithm to divide multiple objects into groups, thereby improving the efficiency of group division. Figure 1 A process 100 according to an embodiment of a data processing method of the present disclosure is shown. The data processing method includes the following steps:

[0020] Step 101: Obtain object attribute data of multiple objects in a target activity scene.

[0021] In this embodiment, object attribute data is used to characterize the object characteristics of each object in the target activity scene. The object attribute data can come from the behavior data accumulated by multiple objects in a historical time period, or from the current behavior data of multiple objects collected in real time. The target activity scene is the activity environment of the object. When the type of object is different, the target activity scene is different.

[0022] In this embodiment, object attribute data is structured data in a specific field. For example, if the target activity scenario is an educational scenario, the object attribute data is obtained by preprocessing student activity data. Activity data includes students' academic performance, learning style (e.g., visual / auditory), collaboration ability rating, project participation history, etc. Activity data is preprocessed (e.g., discretization and normalization) to convert it into numerical features that can be processed by the algorithm, ensuring that attributes of different dimensions (e.g., categorical and continuous variables) are appropriately weighted when grouping.

[0023] For example, if the target activity scenario is a medical setting, the object attribute data is derived from preprocessed patient medical data, including physiological indicators (such as blood pressure and blood sugar), medical history, and treatment response records. Genetic algorithms must incorporate the characteristics of medical data (such as time series and multimodal data) to optimize the heterogeneity of treatment groups and ensure complementarity among group members in their treatment plans.

[0024] For example, if the target activity scenario is an industrial one, the object attribute data is pre-processed from the equipment's production data. This production data includes equipment operating parameters (such as failure rates and maintenance cycles) and production line task requirements. Using a genetic algorithm, we dynamically adjust equipment grouping to optimize resource allocation efficiency and reduce downtime.

[0025] In each of the aforementioned scenarios, object attribute data is not abstract numerical values but rather structured information closely tied to domain knowledge. This information must be converted into a coded form that can be processed by genetic algorithms through technical means (such as feature engineering and data standardization). For example, in educational scenarios, learning styles must be mapped to discrete numerical values, and normalized to eliminate dimensional differences, ensuring that the algorithm can accurately measure the collaborative potential of students when grouping them.

[0026] Optionally, the above step 101 includes: obtaining initial collected data of multiple objects in the target activity scene, wherein the initial collected data can be historical data (such as students' past learning behavior records, employees' historical collaboration data, etc.) or real-time generated interactive data (such as students' instant interactive features in online classes), and the initial collected data is characterized to obtain object attribute data.

[0027] Optionally, a numerical discretization operation is performed on the object attribute data. The discretization operation can normalize all object attribute data to ensure that the data is uniformly mapped to the same range, thereby avoiding deviations when calculating the fitness function. The following is the formula for scaling attribute values to the range of 0 to 1:

[0028] a norm =(aa min ) / (a max -a min ) (1)

[0029] In formula (1), a norm belongs to attribute a i The normalized result of the value of a min and a max They are the minimum and maximum values of the corresponding object attribute data respectively.

[0030] The collection, storage, use, processing, transmission, provision, and disclosure of object attribute data in this disclosure comply with relevant laws and regulations and do not violate public order and good morals. User-related information in object attribute data is collected and processed with the user's consent and in a confidential manner.

[0031] Step 102: Obtain the operator body and configuration parameters of the algorithm operator.

[0032] In this embodiment, an operator is a mapping or calculation rule. In mathematics, an operator maps elements in one set to another in a specific way. The operator body is the specific content of the operator. An algorithm operator can have multiple operator bodies, each of which can belong to an operator type such as selection, crossover, or mutation in a genetic algorithm. Configuration parameters are parameters after configuring the operators in the algorithm. For example, configuration parameters include the execution sequence number of the operator body, the priority of the operator body, the position of the operator body, and the number of executions.

[0033] In this embodiment, the operator ontology allows the execution subject on which the data processing method runs to determine the specific operator. Configuration parameters allow the execution subject to determine the position, priority, and number of times each operator ontology appears in the entire genetic algorithm. The operator ontology can be the same operator as or different from the operator in a traditional genetic algorithm. Configuration parameters can be parameters set by the user for the operator ontology in the genetic algorithm based on the requirements of the target activity scenario.

[0034] Step 103: Determine a genetic algorithm based on the operator ontology and configuration parameters.

[0035] In this example, a genetic algorithm encodes potential solutions to a problem using a chromosome-like data structure composed of genes. It then applies various recombination operators to generate optimal solutions based on a cost function. Note that a chromosome represents a possible solution to a particular problem, a gene is a unit on a chromosome, and an allele is a specific value of a gene on a particular chromosome.

[0036] The genetic algorithm disclosed in the present invention is a genetic algorithm generated based on operator ontology and configuration parameters. Compared with traditional genetic algorithms, the operator ontology of the genetic algorithm disclosed in the present invention can be designed independently, and the configuration parameters can be set based on demand. The genetic algorithm disclosed in the present invention is more flexible than traditional genetic algorithms. Specifically, traditional genetic algorithms often use fixed operators, which are difficult to adapt to the complex needs of different fields. For example, educational scenarios may require highly heterogeneous grouping to promote collaboration, while industrial scenarios may pursue homogeneity to improve efficiency. The present invention enables genetic algorithms to adapt to various scenarios by providing a variety of operator ontologies (including innovative operators and classic operators) and configurable configuration parameters. By flexibly selecting operator ontologies and configuration parameters, users can not only optimize the quality of results for specific problems (such as selecting "group crossover + tournament selection" for rapid convergence), but also balance efficiency and accuracy (such as reducing the number of operators to shorten running time).

[0037] Step 104 : Based on the object attribute data, a genetic algorithm is used to divide the multiple objects into groups to obtain a group division result.

[0038] In this embodiment, the execution entity running the data processing method encodes potential solutions to problems using a chromosome-like data structure for object attribute data. This structure is composed of genes and generates an optimal solution based on a cost function using multiple recombination operators. It is worth noting that a chromosome represents a possible solution to a particular problem, a gene is a unit on a chromosome, and an allele is a specific value of a gene on a specific chromosome.

[0039] In this embodiment, the genetic algorithm performs a series of genetic operations in a probabilistic manner, including elite retention operations, selection operations, crossover operations, and mutation operations, wherein the operators in the above-mentioned selection operations, crossover operations, and mutation operations are operations obtained after being improved under the guidance of the operator body and configuration parameters.

[0040] In this embodiment, the elite retention operation passes the best chromosomes in the current population corresponding to the elitism percentage to the next generation, ensuring that those chromosomes that are highly adapted to the environment are retained during the reproduction process.

[0041] In this example, selection plays a crucial role in population evolution. The operators in this selection operation follow the "survival of the fittest" principle and are based on the theoretical assumption that individuals with the highest fitness produce better offspring. Therefore, the selection operator employs various techniques to select the most suitable chromosomes, ensuring that they, or portions of them, survive in the next generation after genetic manipulation.

[0042] In this embodiment, the crossover operation is crucial for exploring the search space, and its purpose is to improve the quality of the new generation population. This operation combines two chromosomes (i.e., parents) selected through the selection process and generates offspring with higher fitness by exchanging their genetic information. The two chromosomes with the highest fitness among the parents and offspring will be selected for the next generation. The execution of the crossover operation is based on the principle that good genes should be retained, and it is performed according to a certain crossover probability.

[0043] In this example, the mutation operation acts on chromosomes with a specific probability, introducing new genetic information by changing a small number of genes, thereby preventing the algorithm from prematurely converging to a local optimum. While various mutation operators exist in the literature, this study employs the well-known crossover mutation, which swaps the positions of two randomly selected genes. This operator was chosen for its simplicity and effectiveness, as well as its ability to avoid inconsistencies within the population, such as populations with unequal membership or duplicate genes.

[0044] In this embodiment, the genetic algorithm will continue to generate new generations until the search space is fully explored, and a variety of different termination conditions are specifically adopted.

[0045] Optionally, the configuration parameters may further include: a maximum number of iterations of the genetic algorithm. For example, the genetic algorithm is designed to automatically terminate when the maximum number of iterations specified by the configuration parameters is reached.

[0046] In this embodiment, the group division results include two levels of structured information. The first is the specific group number to which each object belongs; the member attributes of each group meet the preset homogeneous / heterogeneous mixing constraints. The group division results can ensure that the members in the group achieve the optimal collaborative configuration in the specified attribute dimension.

[0047] Step 105 : sending grouping information to the multiple objects based on the group division result, so that the multiple objects can move in the target activity scene based on the grouping information.

[0048] In this embodiment, the group division result is a group division result of multiple object groups given by the genetic algorithm, such as Figure 2 As shown, the plurality of objects is 12 objects. Specifically, the object IDs of the 12 objects are: 2, 4, 7, 10, 8, 3, 1, 12, 9, 5, 6, 11. The 12 objects can be divided into three groups g1, g2, and g3 using the genetic algorithm. In these groups, each object has a corresponding position index, such as the object in group 3 is at position 5.

[0049] like Figure 2As shown, each solution (also called chromosome) will be converted into an array whose length is equal to the total number of objects m. Each element in the array (called gene) represents an object ID, and the position (index) of the element indicates the group to which the object should be assigned. Figure 2 In the equation, w refers to the position of object 3 in 5, and is assigned to group g2. The group of the object is calculated by formula (2):

[0050] (object_index div group_size)+1=(5 div 4)+1=2→g2 (2)

[0051] In formula (2), object_index is the position of the element, and group_size is the number of elements in the group.

[0052] In this embodiment, the initial population is randomly generated based on the adopted encoding scheme. It should be noted that this population consists of a certain number of chromosomes, which are actually feasible encoding solutions. The total number of chromosomes matches the preset population size. The purpose of using a random generation strategy is to increase the diversity of the population and strengthen the algorithm's ability to converge to the optimal solution.

[0053] The data processing method provided by the embodiments of the present disclosure first obtains object attribute data of multiple objects in a target activity scene, where the object attribute data is used to characterize the object characteristics of each object in the target activity scene; secondly, obtains the operator ontology and configuration parameters of the user algorithm operator; thirdly, determines a genetic algorithm based on the operator ontology and configuration parameters; thirdly, uses the genetic algorithm to group the multiple objects based on the object attribute data to obtain a group division result; and finally, based on the group division result, sends grouping information to the multiple objects so that the multiple objects are active in the target activity scene based on the grouping information. Thus, through the user's parameter setting of the algorithm operator of the genetic algorithm, the genetic algorithm is made to flexibly adapt to different types of object attribute data, thereby improving the genetic algorithm's ability to perform comprehensive analysis on multiple factor data, improving the generalization ability and adaptability of the genetic algorithm, and improving the efficiency of group division.

[0054] In some optional implementations of the present disclosure, the above-mentioned method of dividing multiple objects into groups based on object attribute data using a genetic algorithm to obtain group division results includes: grouping multiple objects into chromosomes based on the object attribute data using a genetic algorithm, and calculating the fitness functions of the chromosomes of the multiple objects; based on the fitness function, detecting whether a population fitness evaluation condition is met; in response to detecting that the population fitness evaluation condition is met, using the chromosome grouping results of the multiple objects as the group division results.

[0055] In this optional implementation, the fitness function evaluates the effectiveness of a chromosome's gene combination, reflecting its corresponding cost function value, thereby guiding the selection of the optimal chromosome. During the optimization process, the fitness function and the cost function are negatively correlated: the lower the cost function value, the higher the fitness function value. Each chromosome has a fitness function, and its value represents the chromosome's ability to "perform." Generally speaking, higher fitness values indicate better chromosomes, which can better solve the input problem.

[0056] In this optional implementation, the group fitness evaluation condition is to meet the conditions for assigning individuals to different groups of specific tasks or goals. Through this condition, these individuals can effectively interact and cooperate within the group to jointly complete specific tasks or goals.

[0057] In this optional implementation, the group formation problem involves a group of n objects O = {o1, o2, o3, ..., o n}, these objects should be grouped into a set of k groups G = {g1, g2, g3, ..., g k}. Each object has z attributes A i ={a i，1 , a i，2 , a i，3 ,...,a i，z}, these attribute values are used to form groups. Each object can only belong to one group, and each group should have m members (group size). Group types can be: a. Heterogeneous groups, where the attribute values of members should be different; b. Homogeneous groups, where the attribute values of members should be the same; c. Mixed groups, where some attribute values are the same and other attribute values are different. The goal of group formation is to generate the optimal group by optimizing the fitness function f(x) based on the group type and the attributes of the objects. The goal of the problem is to minimize this function, because the lower the value of the cost function, the higher the fitness of the group. The optimal group f(x) = 0.

[0058] The disclosed method proceeds to calculate the fitness of each chromosome. This step is achieved by evaluating the fitness value of each group formed by each chromosome. The fitness function is closely related to the appropriateness of group formation, and its core goal is to reduce the total variance.

[0059] The disclosed method uses Euclidean distance as the metric for the fitness function, which is a basic distance metric widely used in genetic algorithms. The system has the ability to create mixed groups, which means pursuing homogeneity in some attributes and heterogeneity in other attributes. Therefore, q and r are set to represent the number of attributes that require homogeneity and heterogeneity, respectively, that is, q+r=z (where z represents the total number of attributes). In the case of homogeneous attributes, objects in the same group should show similar values on these features. Based on this, two objects o i and o j The similarity on q attributes can be calculated using the following formula (3):

[0060]

[0061] In formula (3), a i，c and a j，c Respectively represent the object o i and o j Attribute c.

[0062] By reducing the distance, the best similarity can be achieved. The homogeneity between objects in a group can be calculated using formula (4):

[0063]

[0064] On the other hand, the so-called heterogeneous attributes mean that objects in the same group should show differences in specific features. Therefore, the difference between two objects in r attributes can be expressed by the function shown in formula (5):

[0065]

[0066] Obviously, the greater the distance, the greater the difference between the values. Therefore, according to formula (5), when the OD function is minimized, the optimal difference can be obtained. The heterogeneity between objects in the group can be expressed by the function shown in formula (6):

[0067]

[0068] Based on the above, the fitness of the group is calculated as follows:

[0069] G Fitness(g t )=G Homo(g t )+G Hete(g t ) (7)

[0070] It is important to note that when no relevant attributes are specified, that is, when the value of q or r is set to zero, the user has the option to create purely homogeneous or heterogeneous groups, rather than mixed groups. Therefore, in this case, the functions GHomo and GHete will not be calculated.

[0071] In this embodiment, the cost function is used to quantify the "cost" of the current solution (such as grouping bias and resource consumption), with lower values representing better solutions. The fitness function is constructed based on the cost function. Through mathematical transformations (such as taking its inverse, negating it, or performing a linear transformation), the cost function can be mapped to the fitness function, thereby obtaining the fitness value of the fitness function.

[0072] In order to construct the optimal mixed group, the goal is to improve the fitness value of the fitness function, that is, to reduce the value of the cost function GFitness. Accordingly, the optimization goal of this grouping problem is to minimize the cost function of the chromosome. The specific expression of the cost function is shown in formula (8):

[0073]

[0074] In formula (8), the lower the value of the cost function CFitness, the better the solution represented by chromosome c.

[0075] It should be noted that, in actual genetic operations, the genetic algorithm will also perform chromosome screening guided by maximizing the corresponding fitness value (such as controlling the fitness function F=-CFitness(c)).

[0076] The method for obtaining group division results provided by this optional implementation adopts a genetic algorithm to group chromosomes of multiple objects based on object attribute data, and calculates the fitness functions of the chromosomes of multiple objects; based on the fitness function, detects whether the population fitness evaluation conditions are met; in response to detecting that the population fitness evaluation conditions are met, the chromosome grouping results of multiple objects are used as group division results, thereby improving the reliability of the group division results.

[0077] Optionally, in response to detecting that the group fitness evaluation condition is not met, continue to use the genetic algorithm to group the chromosomes of multiple objects, and calculate the fitness function of the chromosomes of multiple objects; based on the fitness function, detect whether the population fitness evaluation condition is met until it is detected that the group fitness evaluation condition is met.

[0078] In some optional implementations of the present disclosure, the above-mentioned operator ontology includes: a selection operator and / or a crossover operator, the configuration parameters include: an execution sequence number of the selection operator and / or the crossover operator, and obtaining the operator ontology and configuration parameters of the algorithm operator includes: determining the selection operator and / or the crossover operator according to the scene information of the target activity scene; sorting the selection operator and / or the crossover operator to obtain the execution sequence number of the selection operator and the crossover operator; based on the operator ontology and the configuration parameters, determining the genetic algorithm includes: sorting the selection operator and / or the crossover operator according to the execution sequence number to obtain the genetic algorithm.

[0079] In this optional implementation, the group division in different scenarios has different requirements for operators. For example, in the learning scenario, the requirements for student group division are relatively low, and the sub-operators of the selection operator of the operator body can be a few. The sub-operators in the selection operator can be sorted based on demand.

[0080] In this optional implementation, the selection operator and crossover operator can be the selection operator and crossover operator in the traditional genetic algorithm. By configuring parameters, the user can configure the selection operator separately, the crossover operator separately, or configure the selection operator and crossover operator at the same time.

[0081] In this optional implementation, the execution sequence number of the selection operator and / or the crossover operator is used to characterize the execution order of the selection operator and / or the crossover operator in the genetic algorithm. Specifically, the execution sequence number of the selection operator and / or the crossover operator includes: the execution sequence number of at least one sub-operator in the selection operator and / or the execution sequence number of at least one sub-operator in the crossover operator. The execution sequence numbers of the selection operator and the crossover operator are sorted in sequence and have a sequential relationship. Compared with the traditional genetic algorithm, by configuring the execution sequence number of the selection operator and / or the crossover operator, the genetic algorithm can be made more adaptable.

[0082] In this optional implementation, the selection operators include: roulette wheel selection, ranking selection, elite selection, and tournament selection. Roulette wheel selection: This is a probability-based selection technique in which the probability of a chromosome being selected is proportional to its fitness. Each chromosome is assigned a probability of selection, which is equal to the ratio of the fitness of that chromosome to the total fitness of the population. The cumulative probability is calculated as the probability of selection of the chromosome plus the sum of the selection probabilities of all previous chromosomes in the population. The sum of the cumulative probabilities is 1, ensuring that every chromosome in the population has a chance to be selected. Subsequently, a random number between 0 and 1 is generated, and the corresponding chromosome is selected for evolution based on which part of the roulette wheel the number falls on. Chromosomes with higher fitness occupy a larger part of the roulette wheel and therefore have a higher probability of being selected.

[0083] Rank selection: This selection method takes into account the ranking of chromosomes in the population when calculating the selection probability. It is similar to roulette wheel selection, but distributes the "pie" more fairly, giving all chromosomes a chance to be selected. This method helps overcome the limitations of roulette wheel selection when chromosome fitness varies greatly. For example, a chromosome with very high fitness can occupy a large part of the roulette wheel, causing it to be selected multiple times, thus reducing the performance of the selection.

[0084] Elite selection: This selection method includes the best chromosomes in the new population. Thus, elite selection creates a pool of elite chromosomes—those with the best fitness values. Because the population contains different chromosomes at each iteration, this pool is redefined. Therefore, the first parent is always chosen from this list, while the other parent is randomly selected from the population. Since one parent is one of the best chromosomes, the probability of generating optimal offspring increases.

[0085] Tournament Selection: In this operator, a "tournament" is held to select parents. Specifically, a tournament size is set, which determines how many chromosomes are randomly selected from the population to compete. The winner, the chromosome with the best fitness, becomes the parent. All chromosomes are returned to the population and have the opportunity to participate in another tournament.

[0086] In this optional implementation, the crossover operators include: one-point crossover and two-point crossover. For a one-point crossover, a crossover point is randomly selected between the two parents, and the genetic material to the right of this point is exchanged. Therefore, each offspring contains the left portion of one parent and the right portion of the other parent, as defined by this point. If the chromosome contains unique genes that should not be repeated, genes in the right portion that do not appear in the left portion are exchanged, and the remaining values are used to fill in the empty genes. A two-point crossover: This operator works similarly to a one-point crossover, except that two crossover points are randomly selected between the two parents, and the genes between these points are exchanged.

[0087] Traditional genetic algorithms typically use a fixed, single selection operator or a limited number of combinations. However, this disclosure flexibly combines different selection and crossover operators (such as innovative operators like gene crossover and group crossover) based on specific scenarios. Experiments have demonstrated the advantages of specific combinations (such as combining ranked selection with group crossover) in improving solution quality and execution efficiency. This dynamic combination strategy is not a conventional approach in existing technologies, but rather an innovation achieved through systematic parameter configuration and optimization.

[0088] The data processing method provided by this optional implementation can effectively adjust the genetic algorithm according to specific application scenarios and needs by providing configuration parameters of flexible operator entities. This flexible genetic algorithm allows users to quickly adapt to different tasks and goals during use, thereby improving the applicability and flexibility of data processing.

[0089] In some optional implementations of the present disclosure, the operator ontology includes: a one-point modified crossover operator, and the one-point modified crossover operator includes: selecting a crossover point from a chromosome group with the lowest chromosome fitness.

[0090] In this optional implementation, chromosomes represent a complete grouping scheme, see Figure 2 For example, if 12 objects are divided into 3 groups, each group is a subgroup divided by the chromosome, with an independent fitness evaluation value (GHomo / GHete calculated by formula 3-6). The above-mentioned "chromosome group with the lowest fitness" refers to the subgroup with the worst group fitness value in the current chromosome (that is, the group with the largest GHomo+GHete value). In this optional implementation, compared with the traditional one-point crossover operator, the intersection point of the one-point modified crossover operator is no longer a random position on the chromosome, but is selected from the group with the lowest fitness. This modification of the point selection strategy may produce higher-quality groups, because at least the worst group will be improved through recombination, while the group with higher fitness will be retained to the next generation.

[0091] The one-point modification crossover operator provided by this optional implementation effectively breaks the local optimum by modifying the crossover operator, so that the fitness function approaches the global optimum more quickly, thereby improving the efficiency of obtaining the group division result.

[0092] In some optional implementations of the present disclosure, the above-mentioned operator body includes: a two-point modified crossover operator, and the two-point modified crossover operator includes: for the two parent chromosome groupings selected by the selection operator, selecting a crossover point from the worst group of one of the two parent chromosome groupings, and selecting another crossover point from the worst group of the other parent chromosome grouping of the two parent chromosome groupings.

[0093] In this optional implementation, the core mechanism of the two-point modified crossover operator is to select a crossover point from the worst population (i.e., the set of chromosomes with the lowest fitness) of each of the two parent chromosomes. The "worst population" here refers to the set of chromosomes with the lowest fitness in the current population, not a single chromosome. This selection can be random or at a specific location.

[0094] The implementation logic of the two-point modified crossover operator can be as follows: In the crossover operation, two parent chromosomes are first selected (i.e., the two chromosomes selected by the selection operator). Then, a crossover point is randomly selected from the "worst population" of each parent chromosome (for example, the gene segments of each parent chromosome that are classified as low fitness). The "one" and "the other" here specifically refer to two different parent chromosomes, not different parts of the same parent. For example, the first crossover point is selected from the worst part of parent chromosome A, and the second crossover point is selected from the worst part of parent chromosome B.

[0095] In this alternative implementation, the two-point modified crossover operator is similar to the one-point modified crossover operator. This operator randomly selects the first crossover point from the worst population of one parent and the second point from the worst population of the other parent. This approach may help prevent the loss of populations with higher fitness while recombining the worst population.

[0096] The two-point modified crossover operator provided by this optional implementation randomly selects the first crossover point from the worst group of one parent and the second point from the worst group of the other parent, which helps prevent the loss of groups with higher fitness and improves the efficiency of group recombination.

[0097] In some optional implementations of the present disclosure, the above-mentioned operator ontology includes: a gene crossover operator, which includes: assigning a random number to each gene, and based on the value of the random number, exchanging the genes in the chromosome grouping so that the exchanged genes meet the population division constraint conditions.

[0098] In this optional implementation, the core innovation of the gene crossover operator is to perform an exchange operation on a single gene through random number judgment. When implementing this operator, although the judgment and exchange are performed on a gene basis, the validity and consistency of the chromosome as a whole must be ensured. Since the genetic algorithm requires that "genes must not be repeated" in the crossover operation, this means that the chromosomes after gene exchange must meet the constraints of group division (for example, each object belongs to only one group). Therefore, the above optional implementation implies the requirement for maintaining the chromosome structure, that is, after gene exchange, technical means (such as conflict detection and repair) must be used to ensure the legitimacy of the chromosome.

[0099] In this alternative implementation, each gene is assigned a random number, similar to the rolling of dice. If the number is even, the parent gene at that position is swapped; otherwise, the gene in the offspring remains unchanged. As with other crossover operators, no genes are repeated after the operation is completed.

[0100] In some optional implementations of the present disclosure, the above-mentioned operator body includes: a group crossover operator, which includes: based on the fitness value, screening out the chromosome grouping set with the worst fitness, selecting each worst part from the parent chromosome grouping, and crossing the worst parts of any two parent chromosome groups.

[0101] In this optional implementation, the above screening to obtain the chromosome grouping set with the worst fitness includes: first, the system will evaluate the fitness of all population groups in the current population (i.e., the grouping scheme corresponding to the chromosome). The fitness value of each population group is calculated by the above formula (6), and the optimization goal is to minimize this value. Subsequently, the system sorts the fitness values of each population group, and the population group with higher fitness value (i.e., higher cost and worse performance) is marked as the "worst population group". For example, if there are 10 population groups in total, the system may select the population groups with the lowest fitness ranking (i.e., the population groups ranked 6th to 10th) as the screening target. Finally, after the screening is completed, for the selected worst population groups, the crossover operation is performed according to the following rules: A) Select the respective worst population groups from the parent chromosomes (e.g., the top 50% population groups with the highest fitness in each parent chromosome). B) Exchange the genes in the worst population groups of the two parents according to a preset ratio (e.g., 50%). For example, the first half of the genes in the worst population group of parent A are recombined with the second half of the genes in the worst population group of parent B to generate a new offspring population group.

[0102] Uniqueness constraint: Genes (i.e., object IDs) must be unique during the exchange process to prevent the same object from being assigned to different groups. If a conflict occurs, the system will automatically fill in the remaining unused genes to maintain grouping integrity.

[0103] The purpose of designing the above rules is to break the local optimal solution by reorganizing the genes of the low-fitness population group and introduce diversity, thereby improving the global search ability of the algorithm.

[0104] In this optional implementation, a group crossover operator swaps genes from the worst-performing groups of parents. Specifically, the number of groups participating in the crossover is first defined, and then the worst-performing group from each parent is selected for crossover. For example, 50% of the chromosome set may be selected for crossover. Using a group crossover operator can expand the problem space, avoid getting stuck in local optima, and introduce new genetic material into the next generation.

[0105] The group crossover operator provided by this optional implementation method selects the chromosome grouping set with the worst fitness based on the fitness value, selects the worst part of each parent chromosome grouping, and crosses the worst parts of any two parent chromosome groups. For these groups, the offspring will include half of the genes of each parent, thereby improving the overall fitness of the group.

[0106] To ensure comparability of experimental results, this disclosure used the same computer (Intel Core i7 12700 / 16GB DDR5 / 512GB SSD) and the same dataset in a practical example. The dataset contained 100 objects and 5 attributes, with the goal of forming a mixed group of 5 members, containing 3 heterogeneous attributes and 2 homogeneous attributes. The genetic parameter values selected for the experiment were: 200 generations, 150 population size, 80% crossover rate, and 5% mutation rate. These parameters were considered reasonable and would not negatively affect the algorithm performance while guiding the algorithm to achieve optimal results.

[0107] Since the grouping genetic algorithm is a metaheuristic method that outputs different solutions each time it is run, each experiment was executed 15 times and the average of the performance measures was used to evaluate the process. This number of executions was chosen based on the literature and was considered a reasonable number to provide reliable results.

[0108] Sensitivity analysis led to extensive experimentation with combinations of selection and crossover operators provided by the system. Regarding selection operators, roulette wheel selection and ranking selection were more time-consuming than elitism and tournament selection. In particular, ranking selection outperformed the other operators, producing solutions with better fitness. This is likely because ranking selection creates greater diversity in the problem space by rebalancing chromosome selection probabilities, allowing for a wider range of chromosomes to be selected. However, since tournament selection also produces appropriate group formations that meet fitness outputs in a shorter time, this selection operator may be a better choice when a fast solution is required, especially when combined with a larger number of crossover operators.

[0109] The results show that the newly introduced gene crossover and group crossover can be considered effective operations, as they are able to produce groups with high fitness within a short execution time. When combining the three crossover operations, significantly better fitness results were observed. In particular, in the current experiment, the combination of the gene crossover operator, the two-point modification crossover operator, and the group crossover operator outperformed other combinations. Furthermore, when combining more crossover operations, the role of the selection operation in the results is relatively small. It is worth noting that using more crossover operations can produce better solutions, but it also significantly increases the completion time. Therefore, users should consider the algorithm's completion time and the quality of the results they desire, depending on the problem they are solving, to select the most appropriate combination of operations that produces the desired results. Despite combining a small number of crossover operations, the algorithm performs very well, providing reliable results.

[0110] Further references Figure 3 As an implementation of the methods shown in the above figures, the present disclosure provides an embodiment of a data processing device, which is similar to Figure 1 Corresponding to the method embodiment shown, the device can be specifically applied to various electronic devices.

[0111] like Figure 3 As shown, the data processing device 300 provided in this embodiment includes: an attribute acquisition unit 301, a parameter acquisition unit 302, a determination unit 303, a division unit 304, and a grouping unit 305. The attribute acquisition unit 301 can be configured to acquire object attribute data of multiple objects in the target activity scene, and the object attribute data is used to characterize the object characteristics of each object in the target activity scene. The parameter acquisition unit 302 can be configured to acquire the operator ontology and configuration parameters of the algorithm operator. The determination unit 303 can be configured to determine the genetic algorithm based on the operator ontology and configuration parameters. The division unit 304 can be configured to use the genetic algorithm to group the multiple objects based on the object attribute data to obtain a group division result. The grouping unit 305 can be configured to send grouping information to the multiple objects based on the group division result, so that the multiple objects can be active in the target activity scene based on the grouping information.

[0112] In this embodiment, the specific processing of the attribute acquisition unit 301, the parameter acquisition unit 302, the determination unit 303, the division unit 304, and the grouping unit 305 and the technical effects thereof can be referred to in the respective Figure 1 The relevant descriptions of step 101, step 102, step 103, step 104 and step 105 in the corresponding embodiment are not repeated here.

[0113] In some embodiments of the present disclosure, the above-mentioned division unit 304 is configured to: based on object attribute data, use a genetic algorithm to group chromosomes of multiple objects, and calculate the fitness function of the chromosomes of multiple objects; based on the fitness function, detect whether the population fitness evaluation condition is met; in response to detecting that the population fitness evaluation condition is met, use the chromosome grouping results of multiple objects as the group division result.

[0114] In some optional implementations of the present disclosure, the algorithm includes a selection operator and / or a crossover operator, and the configuration parameters include an execution sequence number of the selection operator and / or crossover operator. The parameter acquisition unit 302 is configured to determine the selection operator and / or crossover operator based on scene information of the target activity scene; and sort the selection operator and / or crossover operator to obtain the execution sequence numbers of the selection operator and crossover operator. The determination unit 303 is configured to sort the selection operator and / or crossover operator according to the execution sequence number to obtain a genetic algorithm.

[0115] In some optional implementations of the present disclosure, the operator ontology includes: a one-point modified crossover operator, and the one-point modified crossover operator includes: selecting a crossover point from a chromosome group with the lowest chromosome fitness.

[0116] In some optional implementations of the present disclosure, the above-mentioned operator ontology includes: a two-point modified crossover operator, and the two-point modified crossover operator includes: for the two parent chromosome groupings selected by the selection operator, selecting one of the worst groups of the two parent chromosome groupings as a crossover point, and selecting another of the worst groups of the other parent chromosome groupings as a crossover point.

[0117] In some optional implementations of the present disclosure, the above-mentioned operator ontology includes: a gene crossover operator, which includes: assigning a random number to each gene, and based on the value of the random number, exchanging the genes in the chromosome grouping so that the exchanged genes meet the population division constraint conditions.

[0118] In some optional implementations of the present disclosure, the above-mentioned operator body includes: a group crossover operator, which includes: based on the fitness value, screening out the chromosome grouping set with the worst fitness, selecting each worst part from the parent chromosome grouping, and crossing the worst parts of any two parent chromosome groups.

[0119] The data processing device provided by the embodiment of the present disclosure includes: first, the attribute acquisition unit 301 acquires object attribute data of multiple objects in the target activity scene, and the object attribute data is used to characterize the object characteristics of each object in the target activity scene; second, the parameter acquisition unit 302 acquires the operator body and configuration parameters of the user algorithm operator; third, the determination unit 303 determines the genetic algorithm based on the operator body and configuration parameters; third, the division unit 304 uses the genetic algorithm to group the multiple objects based on the object attribute data to obtain the group division results; finally, the grouping unit 305 sends grouping information to the multiple objects based on the group division results, so that the multiple objects are active in the target activity scene based on the grouping information. Thus, through the user's parameter setting of the algorithm operator of the genetic algorithm, the genetic algorithm can flexibly adapt to different types of object attribute data, improve the genetic algorithm's ability to perform comprehensive analysis on multiple factor data, improve the generalization ability and adaptability of the genetic algorithm, and improve the efficiency of group division.

[0120] According to an embodiment of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium, and a computer program product.

[0121] Figure 4A schematic block diagram of an example electronic device 400 that can be used to implement an embodiment of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital assistants, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their modes are provided for example only and are not intended to limit the implementation of the present disclosure described and / or claimed herein.

[0122] like Figure 4 As shown, the device 400 includes a computing unit 401, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 402 or a computer program loaded from a storage unit 408 into a random access memory (RAM) 403. Various programs and data required for the operation of the device 400 can also be stored in the RAM 403. The computing unit 401, the ROM 402, and the RAM 403 are connected to each other via a bus 404. An input / output (I / O) interface 405 is also connected to the bus 404.

[0123] Various components in device 400 are connected to I / O interface 405, including an input unit 406, such as a keyboard, mouse, etc.; an output unit 407, such as various types of displays, speakers, etc.; a storage unit 408, such as a magnetic disk, optical disk, etc.; and a communication unit 409, such as a network card, modem, wireless communication transceiver, etc. Communication unit 409 allows device 400 to exchange information / data with other devices via a computer network such as the Internet and / or various telecommunication networks.

[0124] The computing unit 401 can be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 401 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units that run machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The computing unit 401 performs the various methods and processes described above, such as the data processing method. For example, in some embodiments, the data processing method can be implemented as a computer software program that is tangibly contained in a machine-readable medium, such as the storage unit 408. In some embodiments, part or all of the computer program can be loaded and / or installed on the device 400 via the ROM 402 and / or the communication unit 409. When the computer program is loaded into the RAM 403 and executed by the computing unit 401, one or more steps of the data processing method described above can be performed. Alternatively, in other embodiments, the computing unit 401 can be configured to perform the data processing method by any other appropriate means (e.g., by means of firmware).

[0125] Various embodiments of the systems and techniques described herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), system-on-chip systems (SOCs), programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include being implemented in one or more computer programs that are executable and / or interpreted on a programmable system that includes at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0126] The program code for implementing the method of the present disclosure can be written in any combination of one or more programming languages. These program codes can be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device so that when the program code is executed by the processor or controller, the modes / operations specified in the flow charts and / or block diagrams are implemented. The program code can be executed entirely on the machine, partially on the machine, as a stand-alone software package, partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0127] In the context of the present disclosure, a machine-readable medium can be a tangible medium that can contain or store a program for use by or in conjunction with an instruction execution system, device or equipment. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium can include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0128] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0129] The systems and techniques described herein can be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network (LAN), a wide area network (WAN), and the Internet.

[0130] It should be understood that the various forms of the processes shown above can be used to reorder, add, or delete steps. For example, the steps described in this disclosure can be performed in parallel, sequentially, or in a different order, as long as the desired results of the technical solutions disclosed in this disclosure can be achieved. This is not limited herein.

[0131] The foregoing descriptions of specific exemplary embodiments of the present disclosure are for purposes of illustration and description. These descriptions are not intended to limit the present disclosure to the precise forms disclosed, and it is apparent that many variations and modifications are possible in light of the foregoing teachings. The exemplary embodiments have been selected and described for the purpose of explaining the specific principles of the present disclosure and their practical application, thereby enabling those skilled in the art to realize and utilize a variety of exemplary embodiments of the present disclosure and various options and modifications. The scope of the present disclosure is intended to be defined by the claims and their equivalents.

Claims

1. A data processing method, comprising: Acquire object attribute data of a plurality of objects in a target activity scene, wherein the object attribute data is used to characterize object features of each object in the target activity scene; Get the operator ontology and configuration parameters of the algorithm operator; Determining a genetic algorithm based on the operator ontology and configuration parameters; Based on the object attribute data, the plurality of objects are divided into groups using the genetic algorithm to obtain a group division result; Based on the group division result, grouping information is sent to the multiple objects, so that the multiple objects are active in the target activity scene based on the grouping information.

2. The method according to claim 1, wherein The grouping of the plurality of objects using the genetic algorithm based on the object attribute data to obtain the grouping result includes: Based on the object attribute data, the genetic algorithm is used to group the multiple objects into chromosomes, and the fitness functions of the chromosomes of the multiple objects are calculated; Based on the fitness function, detecting whether a population fitness evaluation condition is satisfied; in response to detecting that the population fitness evaluation condition is satisfied, taking the chromosome grouping results of the plurality of objects as a population division result.

3. The method according to claim 1, wherein The operator ontology includes: a selection operator and / or a crossover operator, the configuration parameters include: an execution sequence number of the selection operator and / or the crossover operator, and the operator ontology and configuration parameters of the acquisition algorithm operator include: Determining a selection operator and / or a crossover operator according to the scene information of the target activity scene; Sorting the selection operator and / or the crossover operator to obtain execution sequence numbers of the selection operator and the crossover operator; Determining the genetic algorithm based on the operator ontology and configuration parameters includes: The selection operator and / or the crossover operator are sorted according to the execution sequence number to obtain a genetic algorithm.

4. The method according to any one of claims 1 to 3, wherein: The operator body includes: a one-point modification crossover operator, and the one-point modification crossover operator includes: selecting a crossover point from a chromosome group with the lowest chromosome fitness.

5. The method according to any one of claims 1 to 3, wherein: The operator body includes: a two-point modified crossover operator, and the two-point modified crossover operator includes: for two parent chromosome groups selected by the selection operator, selecting a crossover point from the worst group of one parent chromosome group of the two parent chromosome groups, and selecting another crossover point from the worst group of the other parent chromosome group of the two parent chromosome groups.

6. The method according to any one of claims 1 to 3, wherein: The operator body includes: a gene crossover operator, which includes: assigning a random number to each gene, and exchanging genes in the chromosome group based on the value of the random number, so that the exchanged genes meet the group division constraint condition.

7. The method according to any one of claims 1 to 3, wherein: The operator body includes: a group crossover operator, which includes: based on the fitness value, screening out the chromosome grouping set with the worst fitness, selecting the worst parts of each parent chromosome grouping, and crossing the worst parts of any two parent chromosome groups.

8. A data processing device, comprising: an attribute acquisition unit configured to acquire object attribute data of a plurality of objects in a target activity scene, wherein the object attribute data is used to characterize object features of each object in the target activity scene; A parameter acquisition unit, configured to acquire an operator ontology and configuration parameters of an algorithm operator; a determining unit configured to determine a genetic algorithm based on the operator ontology and configuration parameters; a dividing unit configured to divide the plurality of objects into groups based on the object attribute data using the genetic algorithm to obtain a group division result; The grouping unit is configured to send grouping information to the multiple objects based on the group division result, so that the multiple objects are active in the target activity scene based on the grouping information.

9. An electronic device, characterized in that: include: at least one processor; and a memory communicatively coupled to the at least one processor; The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium storing computer instructions, characterized in that: The computer instructions are used to enable the computer to execute the method according to any one of claims 1 to 7.

Images