Intelligent improvement method for collaborative aggregation of industrial Internet business process behaviors

Abstract

The present invention provides an intelligent improvement method for the collaborative aggregation of industrial Internet business process behaviors, including the following steps: Step 1: Data collection. Collect accessory data and production data related to accessories from outsourced manufacturers. Step 2: Data preprocessing. After data cleaning, normalize the data to unify the order of magnitude of the data and remove the dimension. Step 3: Horizontal multi-indicator evaluation. For each type of accessory and outsourced manufacturer's collaborative one-to-many relationship, perform a horizontal multi-indicator evaluation. Step 4: Vertical collaborative aggregation. Driven by demand, perform vertical collaborative aggregation for all accessories and outsourced manufacturers' collaborative many-to-many relationships. Step 5: Confirm the collaborative relationship. According to the collaborative strategy determined in step 4, confirm the collaborative relationship with the outsourced manufacturer and enter the business process related data. The method of the present invention can adaptively and iteratively solve complex optimization problems, determine the collaborative aggregation strategy, and achieve intelligent improvement of the collaborative aggregation of business process behaviors.

Description

Technical Field

[0001] The present invention relates to the field of industrial Internet of Things, and in particular to an intelligent improvement method for collaborative aggregation of industrial Internet business process behaviors. Background Art

[0002] The Industrial Internet (IIoT) is the application of next-generation information and communication technologies to the industrial sector. It closely connects manufacturers, suppliers, customers, products, equipment, and other key elements of industrial operations, transforming all types of information into traversable data. Currently, collecting, aggregating, and analyzing data, unlocking its value, and enabling efficient decision-making are powerful tools for reducing costs and increasing efficiency in the industrial sector.

[0003] The Industrial Internet must consider diverse needs, complex data sources, and difficulties in process collaboration. To strengthen the development of the Industrial Internet and further integrate its applications and technological innovation, it is crucial to eliminate barriers to collaboration across various elements of the business process and enhance the ability to collaboratively aggregate business processes and behaviors. Within the business process, it is necessary to integrate information technology with industrial models to promote enterprise digital transformation, break down barriers to collaboration, and achieve intelligent collaborative aggregation of business processes and behaviors.

[0004] In response to this, many industrial Internet platforms have begun to design collaborative services. For example, Alibaba Cloud has proposed integrating its industrial Internet platform with the enterprise ERP (Enterprise Resource Planning) system integration synchronizes basic data and business data from ERP to the Industrial Internet platform, enabling widespread application of relevant data in the collaborative aggregation of business processes and behaviors. However, during this collaborative aggregation process, many collaborative services only consider the multi-metric requirements of individual products, failing to account for the diverse requirements of enterprises across diverse product categories for the overall collaborative network. This lacks targeted functionality for specific businesses. For example, industries like the construction machinery manufacturing industry, with low production concentration and significant product and scale variations, have a large number of product parts, requiring collaborative production from multiple parties. For example, since the hydraulic transmission of an excavator is connected by oil pipes, the relative positions of the excavator's various components are not restricted by transmission relationships, allowing for greater flexibility in part selection. At the same time, enterprises have varying requirements for the scale of the collaborative network encompassing all parts suppliers. Some enterprises seek to minimize the scale of the collaborative network, ensuring minimal collaboration with as few suppliers as possible to facilitate business operations, while others seek to maximize the scale of the collaborative network, aiming to maximize collaboration with as many suppliers as possible to avoid monopoly operations. How can we achieve intelligent improvements in the collaborative aggregation of business process behaviors for such specific business processes while meeting the enterprise's multi-metric requirements and considering the overall scale of the collaborative network? Existing technologies lack effective approaches to address this issue. Summary of the Invention

[0005] Purpose of the invention: The technical problem to be solved by the present invention is to address the shortcomings of the existing technology. Based on the multi-index evaluation of outsourced manufacturers and the different tendencies of enterprises towards the scale of collaborative networks, the present invention extracts the optimization goal of intelligent improvement of collaborative aggregation of business process behaviors, and then combines the horizontal multi-index evaluation to perform vertical collaborative aggregation. To address this problem, an improved metaheuristic algorithm is used to iteratively search for the optimal solution to determine the collaborative strategy.

[0006] The present invention specifically provides an intelligent enhancement method for collaborative aggregation of industrial Internet business process behaviors, comprising the following steps:

[0007] Step 1: Data collection: Collect parts data and production data related to parts from subcontractors;

[0008] Step 2, data preprocessing: perform data cleaning, normalize the data after cleaning, unify the order of magnitude of the data, and remove the dimension;

[0009] Step 3: Horizontal multi-index evaluation: Conduct a horizontal multi-index evaluation of the one-to-many relationship between each type of accessory and its outsourced manufacturers.

[0010] Step 4: Vertical Collaboration Aggregation: Driven by demand, vertical collaboration aggregation is performed on all the many-to-many relationships between parts and outsourced manufacturers to determine the collaboration strategy.

[0011] Step 5: Confirm the collaborative relationship: Based on the collaborative strategy determined in step 4, confirm the collaborative relationship with the outsourced manufacturer and enter the relevant data of the business process.

[0012] In step 1, the parts data provided by the enterprise to the platform is structured and qualitative. Data from outsourced manufacturers comes from a wider range of sources, including platform databases, parts trading websites, and other third-party sources. This data is categorized into qualitative and quantitative data.

[0013] In step 1, let the set of M parts that need to be outsourced be X = {x1, x2, ..., x M}, x M It represents the Mth type of parts that need to be outsourced, and the set of N outsourcing manufacturers is recorded as Y = {y1,y2,...,y N},y N Represents the Nth outsourced manufacturer, where M and N are both natural numbers.

[0014] Parts data includes parts type, name, demand, and interface specifications. Data sources include internal business data or information systems. Parts-related production data from subcontractors includes available parts types, total production costs, lead times, and historical transaction volumes. Data sources include internal business data or external third-party commercial partners.

[0015] In step 2, there may be invalid or erroneous noise data in the external data set, and data cleaning is required during collection. Common sampling distributions such as t distribution are not suitable for small sample data of non-normal populations. Therefore, the present invention identifies noise data based on Chebyshev's theorem in statistical methods and performs data cleaning. First, the mean μ and standard deviation σ of the attribute value are calculated. According to Chebyshev's theorem, the confidence interval of the population mean probability P is The confidence interval is used to determine whether the data is noise data. If the data is outside the confidence interval, it is determined to be noise data and then deleted.

[0016] In step 2, data covered by existing national standards can be directly compared with the national standards. Data not covered by national standards, such as accessory prices, lead times, and historical transaction volumes, require normalization for subsequent horizontal multi-index evaluation. All data with different attributes are dimensioned, scaled, and mapped to the specified interval. The normalization in step 2 uses a linear function:

[0017]

[0018] Among them, d is the normalized attribute value, d max is the maximum value of the attribute in the original data, d min is the minimum value of the attribute in the original data, d ori is the original data attribute value. ori ∈[0,∞), d∈[0,1].

[0019] In step 3, considering the one-to-many relationship between each type of accessory and the outsourced manufacturer, when conducting horizontal multi-index evaluation, for the i-th type of accessory x i , i∈{1,2,…,M}, there are n i Outsourcing manufacturers can jointly produce this accessory x i , this i The outsourced manufacturer is called accessory x i Collaborative outsourcing manufacturers, accessories x i The set of cooperative outsourcing manufacturers is denoted as Y i , where the jth can cooperate with the outsourcing manufacturer y j For accessories x i The total production cost, delivery cycle, historical transaction volume and other t attribute values ​​participating in the horizontal evaluation are normalized into the index set: y j ∈Y i , in the set Represents the t-th attribute value after normalization, such as Can represent outsourced manufacturers j Production accessoriesx i Total production costs, Can represent outsourced manufacturers j Production accessoriesx i The supply cycle, Can represent outsourced manufacturers j Accessories x i Historical transaction volume, the horizontal multi-index evaluation score of each accessory's collaborative outsourcing manufacturers is calculated by weighting t attribute values:

[0020]

[0021] For accessories x i , outsourced manufacturers j The horizontal multi-index evaluation score, α k is the weight of the kth indicator, k∈{1,2,…,t}, and can be positive or negative depending on the indicator's meaning. For example, lower total production costs correlate with higher multi-indicator evaluation scores, resulting in a negative weight; higher historical trading volume correlates with higher multi-indicator evaluation scores, resulting in a positive weight. The absolute value of the weight is used to adjust the proportion of each indicator, that is, the importance of the indicator, and can be adjusted based on the company's needs.

[0022] In step 4, after a horizontal multi-metric evaluation, each component type has a maximum set of N collaborating outsourced manufacturers, each of which has a multi-metric evaluation score. Building on this horizontal one-to-many collaborative aggregation, this invention, driven by demand, conducts vertical collaborative aggregation, targeting the many-to-many relationships between all components and outsourced manufacturers.

[0023] Enterprises have two different tendencies in business process behavior collaboration: minimizing the size of the collaborative network and maximizing the size of the collaborative network. The former, for the convenience of business development, prefers to collaborate with as few outsourced vendors as possible, with multiple types of parts provided by the same vendor as much as possible. The latter, to avoid monopoly, prefers to collaborate with as many outsourced vendors as possible, with each type of part provided by a different vendor as much as possible. Step 4 uses an improved adaptive genetic algorithm to determine the collaboration strategy, which specifically includes:

[0024] Step 4-1: The collaborative decision between the parts set and the outsourcing manufacturers is determined by the M×N matrix O=[o ij ] M×N Indicates that the element o in the i-th row and j-th column of the matrix ij Indicates accessories x i Is it provided by outsourcing manufacturer y j Production:

[0025]

[0026] The collaborative network scale value s is determined by the number of outsourced manufacturers involved in the collaboration of all accessories in the accessory set, that is, the number of outsourced manufacturers selected according to the collaborative strategy matrix O. The formula is:

[0027]

[0028] Among them, sigh represents the sign function, which is used to obtain the sign of the number. when when

[0029] Combining the horizontal multi-index evaluation, the vertical aggregation and collaborative intelligent improvement optimization goal is expressed as:

[0030]

[0031] Where maximize is a maximization function, which means finding a collaborative strategy O=[o ij ] M×N , so that the optimization target G is maximized, and the value of the optimization target G is calculated by the equation in the above formula. v is the multi-index evaluation score, ω s is the weight of the collaborative network size. If the enterprise wants to minimize the size of the collaborative network, then ω s is a negative value; if the enterprise wants to maximize the scale of the collaborative network, then ω s is a positive value. Driven by demand, it realizes intelligent improvement of behavioral collaborative aggregation. In this formula, st means "limited by", followed by the four constraints of the G maximization function: C1 collaborative strategy ensures that the selected outsourcing manufacturers of the accessories must belong to the set of its collaborative outsourcing manufacturers; C2 ensures that each accessory must select a collaborative outsourcing manufacturer; C4 and C4 constraints determine the weight range;

[0032] Step 4-2, using the improved adaptive genetic algorithm, the collaborative strategy is mapped to the chromosome individual in the algorithm, and the chromosome is encoded using the binary encoding method. Each chromosome individual is represented as a binary string. The collaborative strategy O = [o ij ] M×N The number of binary strings W mapped is:

[0033]

[0034] Set the population size to R, initialize and randomly generate R chromosome individuals; the chromosome population is recorded as CH = {ch1, ch2, ..., ch R}, where ch rDenote the r-th chromosome individual, which are all binary strings, denoted as ch r ={b1,b2,…,b W}, where b W represents the W-th binary bit, taking values of 0 or 1, and r = {1,2,…,R}. Each chromosome individual can be mapped to a collaborative strategy;

[0035] Step 4-3, according to the optimization objective in Step 4-1, adopt a fitness function. The individual fitness of each chromosome is calculated by the collaborative strategy it maps to. The fitness function F(ch r ) is:

[0036]

[0037] The selection operation selects the chromosome individuals in the population that can be inherited to the next generation according to the fitness, that is, screens the collaborative strategies according to the optimization objective. Specifically, the present invention adopts a tournament selection operator. First, randomly select θ individuals from the population, where θ < R, and then select the individual with the highest fitness among them. If the population fitness of the n-th generation follows a normal distribution, the probability p(f = max(f1,f2,...,f θ )) of an individual with fitness f being selected is:

[0038] p(f = max(f1,f2,...,f θ )) = θ·P(F < f) θ-1 ·p(f)

[0039] where, f θ represents the fitness of the θ-th individual randomly selected from the population; θ represents the number of randomly selected individuals, and each f with a subscript represents the fitness of an individual randomly selected from the population. This formula represents the probability that an individual with fitness f appears together with θ - 1 chromosome individuals with lower fitness scores. This operator maintains the population diversity and is not easily trapped in local optimal solutions.

[0040] Step 4-4, the crossover operation performs partial bit exchange on the chromosomes, and the mutation operation randomly changes some bits of the chromosomes. Adaptive crossover and mutation generate new collaborative strategies:

[0041] In traditional genetic algorithms, the crossover probability and mutation probability of the chromosome population remain unchanged, making it impossible to quickly search for the global optimal solution. This is because in the early stages of iteration, larger crossover and mutation probabilities are needed to expand the search range of the optimized solution and facilitate finding the global optimal solution; in the later stages of iteration, smaller crossover and mutation probabilities are needed to preserve the better solution and avoid generating poorer solutions after too many mutations. In order to quickly find an optimized collaborative strategy, the present invention adopts adaptive crossover and mutation, uses the maximum fitness and average fitness of the population to calculate and determine whether the population is evolving towards the optimized solution, and quickly searches for the global optimal solution by adaptively changing the crossover probability and mutation probability, i.e., an optimized collaborative strategy.

[0042] Set the crossover probability range and mutation probability range, and the minimum crossover probability Maximum crossover probability Minimum mutation probability and the maximum mutation probability The maximum fitness of the population in history is recorded as f max , the current average fitness of the population is f avg .

[0043] The crossover operation uses the k-point crossover operator. First, multiple groups of parents are randomly selected. Each group of parents includes two chromosome individuals. The larger fitness of the two chromosome individuals is recorded as the parent fitness f p The parent code string is randomly set with k crossover points. When the last operation is performed, the crossover probability P cor Exchange the genes at the crossover point in the parental individuals to produce new individuals;

[0044]

[0045] The mutation operation uses the bit reversal mutation operator to generate new individuals. First, multiple parent individuals are randomly selected. Then, for each parent individual, the mutation probability P is calculated according to the individual fitness f. mut , perform the inversion mutation operation to generate a new individual: b w The value of Indicates exclusive OR operation, when b w =1, When b w =0, w∈{1,2,…,W}:

[0046]

[0047] Step 4-5 records the individual with the highest fitness in the population, which is the currently optimal cooperative strategy. Determine whether the number of iterations has reached a preset value, which is determined by the population size and the probability of crossover and mutation operations, and is generally set between 100 and 500. If it has not reached the preset value, continue with steps 4-3 and 4-4 to iteratively search for the optimal solution. Otherwise, decode the chromosome individual with the highest fitness and adopt its mapped cooperative strategy.

[0048] In step 5, according to the collaborative strategy determined in step 4, the outsourcing manufacturer y j Send order, if production conditions change, outsourcing manufacturer y j If the order is rejected, the set of cooperative manufacturers is modified, and step 4 is executed again to continue using the improved adaptive genetic algorithm to find the manufacturer y that is not selected for the cooperative manufacturer. j Other optimal solutions; if the outsourcing manufacturer y j When receiving an order, the collaborative relationship is confirmed and relevant data on the business process such as total production costs and delivery cycle are entered.

[0049] Based on the Industrial Internet, this invention proposes a method for intelligently enhancing the collaborative aggregation of business process behaviors, driven by demand, supported by algorithms, and guided by optimization. Focusing on specific industrial scenarios with low industry production concentration and large differences in manufacturers' product offerings and scale, this invention extracts optimization goals for intelligently enhancing the collaborative aggregation of business process behaviors based on multi-metric evaluations of outsourced manufacturers and the different preferences of enterprises for collaborative network scale. This method improves metaheuristic algorithms based on computational intelligence mechanisms, adaptively and iteratively solves complex optimization problems, and ultimately achieves intelligent enhancement of the collaborative aggregation of business process behaviors.

[0050] Compared with the prior art, the present invention has the following beneficial effects:

[0051] (1) Driven by demand, focusing on specific industrial scenarios with low industry production concentration and large differences in manufacturers' operating products and scale, based on the multi-index evaluation of outsourced manufacturers and the different tendencies of enterprises in the scale of collaborative networks in business process behavior collaboration, an intelligent improvement method for the collaborative aggregation of industrial Internet business process behaviors is proposed.

[0052] (2) Extract the optimization goals for intelligent improvement of collaborative aggregation of business process behaviors, improve the metaheuristic algorithm based on the mechanism of computational intelligence, adaptively and iteratively solve complex optimization problems, determine the collaborative aggregation strategy, and realize intelligent improvement of collaborative aggregation of business process behaviors. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments, and the above and / or other advantages of the present invention will become more apparent.

[0054] Figure 1 It is a flow chart of the steps of the present invention.

[0055] Figure 2 It is a business process diagram involved in the present invention. DETAILED DESCRIPTION

[0056] like Figure 1 As shown, the present invention provides an intelligent improvement method for collaborative aggregation of industrial Internet business process behaviors, including the following steps:

[0057] Step 1: Collect data. Collect parts data and production data related to parts from subcontractors.

[0058] Taking the rotary reducer of a hydraulic excavator as an example, 50 rotary reducers of hydraulic excavators need to be produced by outsourcing, and the interface of the parts all use national standard data. The set of 14 parts that need to be outsourced is recorded as X = {x1, x2, ..., x 14}, the set of 16 outsourcing manufacturers is denoted as Y = {y1,y2,...,y 16 The accessory serial number, material code, accessory name and required quantity are shown in Table 1. The accessories that each subcontractor can produce are shown in Table 2.

[0059] Table 1

[0060] Accessory number Material Code Accessory name Number of accessories 1 860126741 oil seal 100 2 860126740 bearings 100 3 803009145 drive shaft 100 4 860139454 Slip-on board 50 5 860126736 Slip-on boots 450 6 860126733 plunger cylinder 50 7 860126745 Oil distribution plate 50 8 860126744 case 50 9 860126757 Anti-reverse valve 100 10 860139455 Relief valve 100 11 860120560 Delay valve 50 12 860120607 Dust seal 50 13 860120595 Spherical roller bearings 100 14 803009312 Elastic retaining ring 50

[0061] Table 2

[0062]

[0063]

[0064] Step 2: Data preprocessing: After data cleaning, normalization is performed to unify the order of magnitude of the data and remove dimensions.

[0065] There may be invalid or erroneous noise data in the external data set, which requires data cleaning during collection. This paper identifies noise data based on the Chebyshev theorem in statistical methods and performs data cleaning. First, the mean μ and standard deviation σ of the attribute value are calculated. According to the Chebyshev theorem, the confidence interval of the population mean probability P is Based on this confidence interval, we determine whether the data is noise data. If the data falls outside the confidence interval, it is considered noise data. For example, six third-party parts websites all have historical transaction records for part x1 from outsourced manufacturer y1. The mean is 2055 units, the standard deviation is 128.16, and the probability of the population mean P is 90%, that is, the confidence interval is [773.4, 3336.6]. Therefore, the data with a historical transaction volume of 500 units has a very large deviation and is therefore considered noise data and removed.

[0066] For data that are already regulated by national standards, they can be directly compared with national standards for judgment. For data that are not regulated by national standards, such as accessory prices, delivery cycles, historical transaction volumes, etc., normalization processing is required for subsequent comparison and evaluation. All different attribute data are dimensioned, scaled proportionally, and mapped to the specified interval. The normalization processing in this invention adopts linear function normalization, that is,

[0067]

[0068] Among them, d is the normalized attribute value, d max is the maximum value of the attribute in the original data, d min is the minimum value of the attribute in the original data, d ori is the original data attribute value. ori ∈[0,∞), and after normalization d∈[0,1]. In this embodiment, for the total production cost of accessory x1, the price of the subcontractor y1 is 2600 yuan, the highest price in the set is 4000 yuan, and the lowest price is 400 yuan. After normalization, the price attribute value of the subcontractor y1 for accessory x1 is a dimensionless value of 0.611.

[0069] Step 3: Horizontal multi-index evaluation: Conduct a horizontal multi-index evaluation of the one-to-many relationship between each type of accessory and its outsourced manufacturers.

[0070] Consider the one-to-many relationship between each type of accessory and its outsourcing manufacturers, that is, for accessory x i (i∈{1,2,…,M}), there are n i Outsourcing manufacturers can jointly produce the accessories, this i The outsourced manufacturer is called accessory x i The set of cooperative outsourcing manufacturers is denoted as Y i For example, the set of collaborating outsourcing manufacturers of accessory x1 is Y i ={y1,y2,y3,y4,y7,y 14 The three attributes participating in the horizontal evaluation, total production cost, lead time, and historical transaction volume, can be collaboratively normalized and weighted to obtain a horizontal multi-index evaluation score. The weights for the three attributes are α1 = -0.35, α2 = -0.30, and α1 = 0.35. The multi-index evaluation scores for this embodiment are shown in Table 3.

[0071] Table 3 Multi-index evaluation scores

[0072]

[0073]

[0074] Step 4: Vertical Collaboration Aggregation: Driven by demand, vertical collaboration aggregation is performed on all the many-to-many relationships between parts and outsourced manufacturers.

[0075] Vertical collaborative aggregation combines the horizontal multi-indicator evaluation in step 3 and considers two different preferences of enterprises for collaborative network scale. It includes the following five parts:

[0076] (1) Demand-driven, formulated optimization goals for collaborative aggregation.

[0077] The collaborative strategy between accessories and outsourcing manufacturers is a subset of the Cartesian product space, which is 16 14 ≈7.2×10 16 A subset of collaborative strategies. Assume that the collaborative decision between the parts set and the outsourcing manufacturers is represented by the matrix O = [o ij ] 14×16 Indicates that the element o in the matrix ij Indicates accessories x i Is it provided by outsourcing manufacturer y j Production, that is

[0078]

[0079] The collaborative network scale value s is determined by the number of outsourced manufacturers involved in the collaboration of all accessories in the accessory set, that is, the number of outsourced manufacturers selected according to the collaborative strategy matrix O, and is calculated by the following formula.

[0080]

[0081] Combined with the horizontal multi-index evaluation, the vertical aggregation and collaborative intelligent improvement optimization goal is expressed as

[0082]

[0083] Among them, ω v is the multi-index evaluation score, ω s is the weight of the collaborative network size. In the embodiment, the enterprise hopes to minimize the collaborative network size and sets ω v =0.8,ω s = -0.2. In this constraint, C1 ensures that the selected outsourcing manufacturer for each component must belong to the set of outsourcing manufacturers that can be coordinated; C2 ensures that each component must select a single outsourcing manufacturer for collaborative production; and C4 and the C4 constraint determine the weight range.

[0084] (2) Use genetic algorithm to initialize the collaborative strategy.

[0085] The collaborative strategy is mapped to the chromosome individual in the algorithm. The present invention adopts binary coding method to encode chromosomes, that is, each chromosome individual is represented as a binary string. Collaborative strategy O = [o ij ]14×16 The number of binary strings mapped is

[0086]

[0087] Set the population size to 100 and initialize 100 randomly generated chromosome individuals. The chromosome population is denoted as CH = {ch1, ch2, ..., ch 100}, where each chromosome individual is a binary string, represented by ch r ={b1,b2,…,b 56}(r={1,2,…,100}), which can be mapped into a collaborative strategy.

[0088] (3) Screen collaborative strategies based on optimization objectives.

[0089] According to the optimization objective in (1), the individual fitness of each chromosome is calculated by its mapped collaborative strategy, and the fitness function is

[0090]

[0091] The selection operation selects chromosome individuals in the population that can be inherited to the next generation based on their fitness, that is, screening collaborative strategies based on the optimization goal. Specifically, the present invention adopts a tournament selection operator, first randomly extracting θ individuals from the population, and then selecting the individual with the highest fitness. In the embodiment, θ is set to 3, that is, 3 chromosome individuals are selected each time, and the two individuals with lower fitness are deleted.

[0092] (4) Generate new cooperation strategies.

[0093] The present invention improves the genetic algorithm, adopts the adaptive crossover operation to exchange some bits of chromosomes, and the mutation operation to randomly change some bits of chromosomes. The purpose of the two operations is to generate new chromosomes, that is, to generate new collaborative strategies.

[0094] Specifically, the crossover operation uses a k-point crossover operator, and in this embodiment, k is set to 28. First, multiple groups of parents are randomly selected, each group of parents includes 2 chromosome individuals, and 28 crossover points are randomly set in the 56-bit encoding string of the parents. Let the minimum crossover probability be Maximum crossover probability If the fitness of two parent individuals is 0.3318 and 0.2481 respectively, the parent fitness is the larger one, 0.3318. When the population's historical maximum fitness is 0.4159 and the current average fitness of the population is 0.2297, the crossover probability is

[0095]

[0096] When operating, the crossover probability P corThe binary codes in the parent individuals were exchanged at 28 crossover points to generate new individuals.

[0097] The mutation operation uses the bit reversal mutation operator to generate new individuals with the minimum mutation probability Maximum mutation probability First, multiple parent individuals are randomly selected, and the mutation probability of each parent individual is calculated based on its fitness. If the parent individual fitness is 0.3318, the mutation probability is

[0098]

[0099] For each parent individual, the mutation probability P mut Perform the reverse mutation operation, i.e. (w∈{1,2,…,56}), generating new individuals.

[0100] (5) Iterative search for the optimal solution, that is, the better collaborative strategy.

[0101] Record the individual with the highest current fitness in the population, which is the currently optimal cooperative strategy. Set the number of iterations to a preset value of 300. If the preset value is not reached, continue to execute (3) (4) and iteratively search for the optimal solution; otherwise, decode the chromosome individual with the highest current fitness and adopt its mapped cooperative strategy.

[0102] Step 5: Confirm the collaborative relationship. Based on the collaborative strategy determined in Step 4, confirm the collaborative relationship with the outsourcing vendor and enter the relevant data for the business process.

[0103] Send an order to the outsourced manufacturer according to the collaborative strategy determined in step 4. If the production conditions change and the outsourced manufacturer rejects the order, modify the set of outsourced manufacturers that can be collaborated with, and then execute step 4 again, continuing to use the improved adaptive genetic algorithm to find other optimization solutions that do not select this manufacturer; if the outsourced manufacturer accepts the order, confirm the collaborative relationship and enter business process-related data such as total production costs and delivery cycle.

[0104] In this example, a total of 14 parts require outsourcing, with 16 potential collaborators. If the scale of the collaborative network is ignored and only the horizontal multi-index evaluation is performed, the collaborative strategy involves 10 outsourced manufacturers ({2, 4, 5, 6, 7, 8, 9, 10, 15, 16}), with an average horizontal multi-index score of 0.7773 for all 14 parts. However, after intelligently enhancing vertical collaborative aggregation and considering both horizontal multi-index evaluation and collaborative network scale, the collaborative strategy involves 4 outsourced manufacturers ({2, 7, 15, 16}), minimizing the scale of the collaborative network. The average multi-index score is 0.6974, a decrease of only 0.0899 compared to the collaborative strategy with 10 outsourced manufacturers.

[0105] In a specific implementation, the present application provides a computer storage medium and a corresponding data processing unit, wherein the computer storage medium is capable of storing a computer program that, when executed by the data processing unit, can run the invention content of the method for intelligently enhancing the collaborative aggregation of industrial Internet business process behaviors provided by the present invention and some or all of the steps in each embodiment. The storage medium can be a magnetic disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM).

[0106] Those skilled in the art can clearly understand that the technical solutions in the embodiments of the present invention can be implemented by means of computer programs and their corresponding general hardware platforms. Based on this understanding, the technical solutions in the embodiments of the present invention, in essence or in other words, the part that contributes to the prior art, can be embodied in the form of a computer program, i.e., a software product. The computer program software product can be stored in a storage medium and includes several instructions for enabling a device including a data processing unit (which can be a personal computer, a server, a single-chip microcomputer, a MUU, or a network device, etc.) to execute the methods described in various embodiments of the present invention or certain parts of the embodiments.

[0107] The present invention provides an intelligent enhancement method for collaboratively aggregating business processes and behaviors on the Industrial Internet. There are many specific methods and approaches for implementing this technical solution. The above is only a preferred embodiment of the present invention. It should be noted that those skilled in the art can make several improvements and modifications without departing from the principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention. Any components not specified in this embodiment can be implemented using existing technologies.

Claims

1. An intelligent enhancement method for collaborative aggregation of industrial Internet business process behaviors, characterized by: The following steps are involved: Step 1: Data collection: Collect parts data and production data related to parts from subcontractors; Step 2, data preprocessing: perform data cleaning and normalization after data cleaning; Step 3: Horizontal multi-index evaluation: Conduct a horizontal multi-index evaluation of the one-to-many relationship between each type of accessory and its outsourced manufacturers. Step 4: Vertical Collaboration Aggregation: For all the many-to-many relationships between components and outsourced manufacturers, perform vertical collaboration aggregation and determine the collaboration strategy. Step 5: Confirm the collaborative relationship: Based on the collaborative strategy determined in step 4, confirm the collaborative relationship with the outsourcing vendor and enter the relevant data of the business process; In step 1, let the set of M parts that need to be outsourced be X = {x1, x2, ..., x M }, x M It represents the Mth type of parts that need to be outsourced, and the set of N outsourcing manufacturers is recorded as Y = {y1,y2,...,y N },y N It represents the Nth outsourcing manufacturer, and both M and N are natural numbers; In step 2, based on Chebyshev’s theorem, we identify the noise data and perform data cleaning. First, we calculate the mean μ and standard deviation σ of the attribute value. According to Chebyshev’s theorem, the confidence interval of the population mean probability P is If the data is outside the confidence interval, it is determined to be noise data and then deleted; In step 2, the normalization process uses linear function normalization: Among them, d is the normalized attribute value, d max is the maximum value of the attribute in the original data, d min is the minimum value of the attribute in the original data, d ori is the original data attribute value, d ori ∈[0,∞), d∈[0,1]; In step 3, for the i-th accessory x that needs to be outsourced i , i∈{1,2,…,M}, there are n i Outsourcing manufacturers can jointly produce this accessory x i , this i The outsourced manufacturer is called accessory x i Collaborative outsourcing manufacturers, accessories x i The set of cooperative outsourcing manufacturers is denoted as Y i , where the jth can cooperate with the outsourcing manufacturer y j For accessories x i The t attribute values ​​participating in the horizontal evaluation, after normalization, the index set is In the collection represents the tth attribute value after normalization. The horizontal multi-index evaluation score of each accessory's collaborative outsourcing manufacturer is calculated by weighting the t attribute values: For accessories x i , outsourced manufacturers j The horizontal multi-index evaluation score, α l is the weight of the kth indicator, k∈{1,2,…,t}; Step 4 includes: Step 4-1: The collaborative decision between the parts set and the outsourcing manufacturers is determined by the M×N matrix O=[o ij ] M×N Indicates that the element o in the i-th row and j-th column of the matrix ij Indicates accessories x i Is it provided by outsourcing manufacturer y j Production: The collaborative network size value s is determined by all the outsourced manufacturers involved in the collaborative work of all the accessories in the accessory set, and the formula is: Among them, sigh represents the sign function, when when when Combining the horizontal multi-index evaluation, the vertical aggregation and collaborative intelligent improvement optimization goal is expressed as: Where maximize is a maximization function, which means finding a collaborative strategy O=[o ij ] M×N , so that the optimization target G is maximized; ω v is the multi-index evaluation score, ω s is the weight of the collaborative network size; st means it is restricted by: Step 4-2: Use the improved adaptive genetic algorithm to map the collaborative strategy to the chromosome individuals in the algorithm, and use the binary encoding method to encode the chromosomes. Each chromosome individual is represented as a binary string. Step 4-3: According to the optimization goal in step 4-1, the fitness function is used, and the individual fitness of each chromosome is calculated by its mapped collaborative strategy; Step 4-4: The crossover operation exchanges some bits of the chromosome, and the mutation operation randomly changes some bits of the chromosome. Adaptive crossover and mutation generate new collaborative strategies. Step 4-5, record the individual with the highest current fitness in the population, that is, the current optimal cooperation strategy: determine whether the number of iterations has reached the preset value; if not, continue to execute steps 4-3 to 4-4, iteratively search for the optimal solution; otherwise, decode the chromosome individual with the highest current fitness and adopt its mapped cooperation strategy.

2. The method according to claim 1, characterized in that In step 4-2, the collaborative strategy O = [o ij ] M×N The number of binary strings W mapped is: Set the population size to R, initialize and randomly generate R chromosome individuals; the chromosome population is recorded as CH = {ch1, ch2, ..., ch R }, where ch r Indicates the rth chromosome individual, denoted as ch r ={b1,b2,…,b W }, b W Represents the Wth binary bit, r={1,2,…,R}, and each chromosome individual can be mapped to a collaborative strategy.

3. The method according to claim 2, characterized in that In step 4-3, the fitness function F(ch r )for: The selection operation selects the chromosomal individuals in the population that can be genetically passed on to the next generation according to fitness, screens the collaborative strategies according to the optimization goal, and uses the tournament selection operator. First, θ individuals are randomly selected from the population, where θ < R, and then the individual with the highest fitness is selected. If the population fitness of the nth generation follows a normal distribution, the probability p(f = max(f1, f2,..., f θ )) is as follows: p(f=max(f1,f2,...,f θ ))=θ·P(F<f) θ-1 ·p(f) Among them, f θ represents the fitness of the θth individual randomly drawn from the population.

4. The method according to claim 3, characterized in that In step 4-4, set the crossover probability range and mutation probability range, and the minimum crossover probability Maximum crossover probability Minimum mutation probability and the maximum mutation probability The maximum fitness of the population in history is recorded as f max , the current average fitness of the population is f avg ; The crossover operation uses the k-point crossover operator. First, multiple groups of parents are randomly selected. Each group of parents includes two chromosome individuals. The larger fitness of the two chromosome individuals is recorded as the parent fitness f p ; K crossover points are randomly set in the parent code string, and the crossover probability P is used for the final operation. cor Exchange the genes at the crossover point in the parent individuals to produce new individuals: The mutation operation uses the bit reversal mutation operator to generate new individuals. First, multiple parent individuals are randomly selected. Then, for each parent individual, the mutation probability P is calculated according to the individual fitness f. mut , perform the inversion mutation operation to generate a new individual: b w The value of Indicates exclusive OR operation, when b w =1, When b w =0, w∈{1,2,…,W}:

5. The method according to claim 4, characterized in that In step 5, orders are sent to the subcontractors according to the collaborative strategy determined in step 4. If production conditions change and the subcontractors reject the orders, the set of collaborating subcontractors is modified and step 4 is repeated, continuing to use the improved adaptive genetic algorithm to find other optimal solutions that do not use subcontractors. If the outsourced manufacturer accepts the order, the collaborative relationship is confirmed and relevant data of the business process is entered.

Images