Industrial value chain optimization method under carbon constraint and storage medium

Abstract

The invention discloses an industrial value chain optimization method under carbon constraint and a storage medium. The method comprises the following steps: constructing a carbon constraint industrial value chain; constructing an optimization objective function of the carbon constraint industrial value chain from the economic dimension, constructing constraint conditions of the carbon constraint industrial value chain from the carbon emission dimension, and constructing an industrial value chain model under the influence of carbon constraint; solving the carbon constraint industry value chain model by a constraint multi-objective optimization algorithm based on evolutionary multi-task processing, and setting a main task and an auxiliary task to carry out knowledge transfer to improve the search capability of a population so as to improve the optimization capability of the algorithm; and solving to obtain an industrial value chain construction optimal scheme under the carbon constraint, and finally completing the construction of the optimal industrial value chain according to the optimal solution, so that the maximum benefit can be obtained under the condition of meeting the carbon constraint.

Description

Technical Field

[0001] The present invention relates to the field of carbon constraint, and particularly relates to a method for optimizing the industrial value chain under carbon constraint and a storage medium. Background Art

[0002] The industrial value chain usually consists of links such as raw material procurement, R & D design, production manufacturing, sales service, etc. The coordination among the above links is an important factor related to the development of the industrial value chain. In the industrial value chain, the chain leader enterprise is in the core position, usually served by the manufacturer enterprise on the chain. In order to achieve the goal of carbon constraint, the chain leader enterprise needs to closely cooperate with the upstream and downstream enterprises of the industrial value chain, and while ensuring product quality, jointly reduce carbon emissions and enhance the competitiveness of the entire industrial value chain. By strengthening cooperation with suppliers and parts manufacturers, the chain leader enterprise can comprehensively improve the competition and cooperation management level of the entire industrial value chain, and thus improve the quality and competitiveness of the products of the industrial value chain. The requirements of carbon constraint are directly related to the energy structure of the enterprise. In the production manufacturing link, the chain leader enterprise can use means such as clean energy technology and energy-saving technology to reduce the carbon emission level, but at the same time, it will have a complex fluctuation effect on product cost and profit. Therefore, it becomes very important for the chain leader enterprise to maximize profits under carbon constraint. Summary of the Invention

[0003] To achieve the above object and other related objects, the present invention discloses a method for optimizing the industrial value chain under carbon constraint, including: Construct an industrial value chain under carbon constraint; Construct an optimization objective function of the industrial value chain under carbon constraint from the economic dimension, construct a constraint condition of the industrial value chain under carbon constraint from the carbon emission dimension, and construct an industrial value chain model under the influence of carbon constraint, where:

[0004] Among them, (U) is the optimization objective equation set, (L) is the constraint objective equation set, f1 is the objective function of the optimization algorithm, x is the decision variable, and W is the constraint condition of the decision variable; f2 is the objective function in the constraint, y is the decision variable, and w is the constraint condition of the decision variable. The carbon emission constraint function is as follows:

[0005] Among them, CZ represents the total carbon footprint of the product, CZ(i) represents the carbon footprint contribution value of each enterprise, and CZ MAX represents the carbon footprint threshold of the product; Solve the industrial value chain model under carbon constraint based on the constraint multi-objective optimization algorithm of evolutionary multi-task processing, set the main task and the auxiliary task to perform knowledge transfer to improve the search ability of the population so as to improve the optimization ability of the algorithm; The optimal solution for constructing the industrial value chain under carbon constraints is obtained, and finally, the optimal industrial value chain is constructed according to the optimal solution.

[0006] Furthermore, the construction of the carbon-constrained industrial value chain includes: Let G = {V, E} represent the industrial value chain, where V = {1, 2,..., B} represents the set of nodal enterprises in the industrial value chain, and B represents the total number of nodes in the industrial value chain; E = {e1,..., ek} represents the set of cooperation relationships between different enterprises in the industrial value chain, including the following steps: At the initial state t = 0 (the moment corresponding to the network state), A initial nodes are generated, and the edges between the nodes are randomly generated; At time t, a new node is added. Different types of newly added candidate nodes are distinguished according to the le value (the category of the node) and the lv value (the module where the node is located, i.e., the division of labor in the industrial value chain). The new edge connection rules are as follows: when le = 1; lv = 1, it means that the newly added node is a node of the first category and the first module, and it can only choose to establish a connection with a node of the second category and the first module; when le = 2, lv = 2, it means that the newly added node is an enterprise node of the second category, and it can choose to establish a connection with enterprise nodes of the first category and the third category, and when connecting, it chooses a node of the second module in the first category and the third category for connection; At the t-th moment, the newly added node selects to connect with A t different existing nodes in the network according to the new edge connection rule. The probability that the newly added node connects with the existing old node i is:

[0007] It increases the degree value of each node but does not affect the preferential connection mechanism, where, represents the degree of node i; represents the total degree of the network at time t; When the number of node data in the model reaches B, the addition of node data stops.

[0008] Furthermore, the constraint conditions for constructing the carbon-constrained industrial value chain from the carbon emission dimension include: The carbon emission is:

[0009] Among them, is the carbon emission; is the consumption of the a-th type of energy; is the average low calorific value of the fuel; is the carbon content per unit calorific value of the fuel; is the oxygen content; 44 / 12 is the conversion coefficient for converting C to CO2; The carbon emission intensity is:

[0010] Among them, is the carbon emission intensity, is the added value of the enterprise; Among them, the processing cost is:

[0011] Among them, represents the processing cost of enterprise i; represents the cost of energy type j of enterprise i; represents the innovation factor data of various energy sources, and the energy types include coal, natural gas, non-fossil energy, and green energy; , represents the carbon emissions corresponding to various energy sources; represents the price data of various energy sources of enterprise i; represents the carbon emission reduction target of enterprise i.

[0012] Furthermore, the constraint conditions for constructing the carbon-constrained industrial value chain from the carbon emission dimension also include: Obtain the overall Gini coefficient of the industrial value chain, including:

[0013] Among them, is the overall Gini coefficient, is the carbon emission of the j-th enterprise in the i-th module; is the carbon emission of the r-th enterprise in the m-th module; Z is the total number of modules; B is the total number of enterprises; is the number of enterprises in the i-th module; is the average carbon emission of the enterprise; Obtain the within-module Gini coefficient and between-module Gini coefficient in the industrial value chain, including:

[0014] Among them, is the within-module Gini coefficient, indicating the degree of carbon emission difference among enterprises within the module; is the between-module Gini coefficient, indicating the degree of carbon emission difference between different modules; is the average carbon emission of enterprises in module i, indicating the average carbon emission of enterprises in module i; is the average carbon emission of enterprises in module m, indicating the average carbon emission of enterprises in module m.

[0015] Furthermore, the constraint conditions for constructing the carbon-constrained industrial value chain from the carbon emission dimension also include: Obtain the overall Gini coefficient of the industrial value chain, including:

[0016] Among them: 、 、 Are the contribution of within-module, between-module, and hypervariable density differences respectively, representing the contribution of carbon emissions differences in different dimensions; Is the relative impact of carbon emissions between the i-th and m-th modules, representing the relative impact between two specified modules; Is the difference in carbon emissions between modules i and m, representing the specific value of the carbon emissions difference between two specified modules; Is the hypervariable first moment.

[0017] Furthermore, the optimization objective function for constructing the carbon-constrained industrial value chain from the economic dimension includes: The economic optimization objective is the profit function of the industrial value chain products, including the product selling price 、Procurement cost 、Manufacturing cost 、Value creation activity cost 、And loss cost , The calculation formula of the industrial value chain profit Is as shown in the following formula:

[0018] Among them, Is the cost for the parts supplier to purchase raw materials, Is the cost for the manufacturer to purchase components; Is the cost of the raw material supplier in the manufacturing process, Is the cost of the parts supplier in the manufacturing and processing process; 、 、 Are the value creation costs of the raw material supplier, parts supplier, and manufacturer respectively; 、 、 Are the loss costs of the raw material supplier, parts supplier, and manufacturer in the operation process respectively.

[0019] Furthermore, it also includes: Create an unconstrained multi-objective optimization problem with zero constraints as a special auxiliary task.

[0020] On the other hand, the present invention also provides a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, the above method is implemented.

[0021] ​By adopting the above technical solution, first construct the carbon-constrained industrial value chain, then construct the optimization objective function of the carbon-constrained industrial value chain from the economic dimension, and construct the constraint conditions of the carbon-constrained industrial value chain from the carbon emission dimension, so as to construct the industrial value chain model under the influence of carbon constraint. Then, use the constraint multi-objective optimization algorithm based on evolutionary multi-task processing to solve the carbon-constrained industrial value chain model, and obtain the optimal solution for the construction of the industrial value chain under carbon constraint, so as to be able to obtain the maximum benefit under the condition of meeting the carbon constraint. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In combination with the accompanying drawings and with reference to the following detailed description, the above and other features, advantages and aspects of the various embodiments of the present disclosure will become more apparent. The drawings are used to better understand the solution and do not constitute a limitation to the present disclosure. In the drawings, the same or similar reference numerals represent the same or similar elements, where: Figure 1 It is a flowchart of an optimization method for an industrial value chain under carbon constraint. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0023] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative efforts shall fall within the protection scope of the present invention.

[0024] Refer to Figure 1 , the embodiments of the present invention provide an optimization method for an industrial value chain under carbon constraint, including the following steps: Construct a carbon-constrained industrial value chain; Construct the optimization objective function of the carbon-constrained industrial value chain from the economic dimension, and construct the constraint conditions of the carbon-constrained industrial value chain from the carbon emission dimension, and construct the industrial value chain model under the influence of carbon constraint, where:

[0025] Among them, (U) is the optimization objective equation set, (L) is the constraint objective equation set, f1 is the objective function of the optimization algorithm, x is the decision variable, and W is the constraint condition of the decision variable; f2 is the objective function in the constraint, y is the decision variable, and w is the constraint condition of the decision variable. The carbon emission constraint function is as follows:

[0026] Among them, CZ represents the total carbon footprint of the product, CZ(i) represents the carbon footprint contribution value of each enterprise, and CZ MAX represents the carbon footprint threshold of the product; Solve the carbon-constrained industrial value chain model with an evolutionary multi-tasking-based constrained multi-objective optimization algorithm. Set the main task and auxiliary tasks for knowledge transfer to improve the search ability of the population and thus enhance the optimization ability of the algorithm; Obtain the optimal solution for constructing the industrial value chain under carbon constraints, and finally complete the construction of the optimal industrial value chain according to the optimal solution.

[0027] Specifically, the industrial value chain includes multiple suppliers, parts manufacturers, and manufacturers. The leading enterprises in the industrial value chain are mainly located downstream of the chain, close to the consumer market, and most of them are manufacturer enterprises responsible for the production of end products. According to the actual situation of the industrial value chain, the industrial value chain of the present invention is an industrial value chain composed of three levels: "supplier - parts manufacturer - manufacturer (chain leader)". In this industrial value chain, the supplier supplies the raw materials required for producing parts to the parts manufacturer, such as ores, metals, rubbers, etc.; the parts manufacturer has the function of processing raw materials and producing product parts; as the last link in product production, the manufacturer mainly provides functions such as assembly, manufacturing, and delivery.

[0028] Constructing the carbon-constrained industrial value chain includes: Let \(G = \{V, E\}\) represent the industrial value chain, where \(V=\{1, 2, \ldots, B\}\) represents the set of node enterprises in the industrial value chain, and \(B\) represents the total number of nodes in the industrial value chain; \(E = \{e_1, \ldots, e_k\}\) represents the set of cooperation relationships between different enterprises in the industrial value chain, including the following steps: At the initial state \(t = 0\) (the moment corresponding to the network state), generate \(A\) initial nodes, and the edges between the nodes are randomly generated; At time \(t\), add a new node. Distinguish the newly added candidate nodes of different categories and different modules according to the \(le\) value (the category of the node) and the \(lv\) value (the module where the node is located, that is, the division of labor in the industrial value chain). The new edge connection rules are as follows: when \(le = 1\); \(lv = 1\), it means that the newly added node is a node of the first category and the first module, and it can only choose to establish a connection with a node of the second category and the first module; when \(le = 2\), \(lv = 2\), it means that the newly added node is an enterprise node of the second category, and it can choose to establish a connection with enterprise nodes of the first category and the third category, and when connecting, choose the nodes of the second module in the first category and the third category for connection; At the \(t\)th moment, the newly added node selects to connect with \(A\) t different existing nodes in the network according to the new edge connection rules. The probability that the newly added node connects with the existing old node \(i\) is:

[0029] It increases the degree value of each node, but does not affect the preferential connection mechanism, where, Denote the degree of node i; Denote the total degree of the network at time t; When the node data in the model reaches B, the addition of node data stops.

[0030] The constraint conditions for constructing the carbon-constrained industrial value chain from the carbon emission dimension include: The carbon emission is:

[0031] Wherein, is the carbon emission; is the consumption of the a-th type of energy; is the average low calorific value of the fuel; is the carbon content per unit calorific value of the fuel; is the oxygen content; 44 / 12 is the conversion coefficient for converting C to CO2; The carbon emission intensity is the carbon emission per unit GDP, which can reflect the intensity of carbon emissions generated by enterprises under the same GDP conditions and is an important indicator for evaluating the low-carbon ability of enterprises. The calculation formula is as follows:

[0032] Wherein, is the carbon emission intensity, is the added value of the enterprise; Wherein, the processing cost is:

[0033] Wherein, represents the processing cost of enterprise i; represents the cost of the j-th type of energy of enterprise i; represents the innovation factor data of various types of energy, and the energy types include coal, natural gas, non-fossil energy, and green energy; , represents the carbon emissions corresponding to various types of energy; represents the price data of various types of energy of enterprise i; represents the carbon emission reduction target of enterprise i.

[0034] In the industrial value network, the carbon emissions of different division modules and different enterprises are different. Studying the differences among them is very important for the optimization of the entire industrial value chain and industrial development. The Dagum Gini coefficient and its decomposition are important methods for studying differences and can solve for enterprises or modules with large carbon emission differences. The present invention uses a carbon emission difference function to divide the sample differences into three parts: inter-module differences, intra-module differences, and hyper-variable density. The overall Gini coefficient reflects the overall differences in carbon emissions from energy consumption of enterprises in the industrial value network, and conducts screening and evaluation of different divisions of the industrial value chain and carbon emissions of different enterprises. The calculation formula is as follows:

[0035] Among them, is the overall Gini coefficient, is the carbon emission of the j-th enterprise in the i-th module; is the carbon emission of the r-th enterprise in the m-th module; Z is the total number of modules; B is the total number of enterprises; is the number of enterprises in the i-th module; is the average carbon emission of the enterprise, is the number of enterprises in the m-th module.

[0036] The Gini coefficient within the module and the Gini coefficient between modules measure the carbon emission differences within and between modules respectively. The specific calculation formulas are as follows:

[0037] Among them, is the Gini coefficient within the module, indicating the degree of carbon emission difference among enterprises within the module; is the Gini coefficient between modules, indicating the degree of carbon emission difference between different modules; is the average carbon emission of enterprises in module i, indicating the average carbon emission of enterprises in module i; is the average carbon emission of enterprises in module m, indicating the average carbon emission of enterprises in module m.

[0038] In order to conduct accurate carbon emission evaluations on each module and each enterprise in the industrial value chain, it is necessary to decompose the overall Gini coefficient to further evaluate the contributions of within-module differences (among enterprises), between-module differences (among modules), and the hypervariance density to the overall difference, so as to obtain the contributions of each part of the industrial value chain to the overall carbon emission. The formula is as follows:

[0039] Among them: , , are the contribution of within-module, between-module, and hypervariance density differences respectively, indicating the contribution of carbon emission differences in different dimensions; is the relative impact of carbon emissions between the i-th and m-th modules, indicating the relative impact between the specified two modules; is the difference in carbon emissions between modules i and m, indicating the specific value of the carbon emission difference between the specified two modules; is the hypervariance first moment, is the proportion of the carbon emission of module i in the overall carbon emission, is the proportion of the carbon emission of module m in the overall carbon emission.

[0040] The optimization objective function for constructing a carbon-constrained industrial value chain from an economic dimension includes: The economic optimization objective is the profit function of the products in the industrial value chain, including the product selling price , procurement cost , manufacturing cost , value creation activity cost , and loss cost . The calculation formula for the profit of the industrial value chain is shown as follows:

[0041] Among them, is the cost for the parts supplier to purchase raw materials, and is the cost for the manufacturer to purchase components; is the cost in the manufacturing process of the raw material supplier, and is the cost in the manufacturing and processing process of the parts supplier; , , are the value creation costs of the raw material supplier, parts supplier, and manufacturer respectively; , , are the loss costs in the operation process of the raw material supplier, parts supplier, and manufacturer respectively.

[0042] Solving the carbon-constrained industrial value chain model using the constrained multi-objective optimization algorithm based on evolutionary multi-task processing specifically includes: An evolutionary multi-task algorithm is used to solve the constrained multi-objective optimization problem constructed by the present invention. In this embodiment, an auxiliary task with simple constraints is created. The constraint condition of the auxiliary task is a constraint subset randomly sampled from all the constraints of the main task. Different constraint subsets may form different feasible regions. Therefore, the created auxiliary task can provide some search biases for the main task, so that more possibilities can be provided for the solution of the main task in the subsequent evolutionary algorithm.

[0043] The embodiment of the present invention also creates a multi-objective optimization problem with zero constraints as a special auxiliary task. Creating an auxiliary task can reduce the search difficulty. Without prior knowledge, the difficulties faced by the constraint subsets included in the auxiliary task are different. In this case, an auxiliary task may not have a complementary evolutionary trial with the main task. In addition, when all the constraints are eliminated, there is no infeasible region that hinders the evolution of the population. In this case, the created unconstrained MOP (i.e., the special auxiliary task) can help the main task cross the infeasible region through the way of knowledge transfer, specifically including: The optimization problem of the present invention is a multi-objective optimization problem including carbon constraints. The carbon constraints in the industrial value chain scenario are complex constraints, including carbon constraints on the products of the industrial value chain and carbon constraints on each enterprise in each division of the chain. It is necessary to constrain and optimize each enterprise on the chain. Traditional constrained multi-objective optimization algorithms have poor optimization capabilities when solving such overall and local constraint problems. Moreover, the carbon constraint of the present invention does not mean minimizing carbon emissions, but maximizing profits as much as possible within the carbon emission limit, that is, only need to meet the constraints within a certain limit. To solve the complex optimization problem, the present invention creates an evolutionary multi-task (EMT) including auxiliary tasks to improve the search ability of the working population. The created auxiliary task can reduce the difficulty of search by searching in the low-dimensional space. For the complex optimization problem of the present invention, the present invention designs an EMT mode to solve the constrained multi-objective optimization problem (CMOP) of the present invention, that is, a constrained multi-objective optimization algorithm based on evolutionary multi-task processing (EMCMO).

[0044] The optimization method of the present invention involves the correlation between two tasks; the two tasks share more identical optimal solutions, and their similarity will be higher in the case of complementary evolutionary experiments. For CMOP, the main difficulty comes from the constraint factors of each objective, which may lead to the generation of several small-scale and disjoint feasible regions. To solve this problem, higher requirements are put forward for the diversity of the population. Therefore, the present invention creates an auxiliary task with simple constraints to improve the utilization rate of some solutions that are feasible for the auxiliary task but not feasible for the main task (CMOP). Generally speaking, the fewer the number of constraints, the fewer difficulties the population encounters. The constraint subset randomly sampled from all constraints can be used as the constraint conditions of the auxiliary task. Different constraint subsets may form different feasible regions; therefore, the created auxiliary task can provide some search biases for the main task. In extreme cases, the set auxiliary task may not contain constraints, making it an unconstrained multi-objective optimization problem (MOP), thus providing more solutions for the main task.

[0045] The present invention creates a "special case" with only two tasks, including a main task and an auxiliary task. In addition, an unconstrained MOP with zero constraints is created as a special auxiliary task, and some practical problems are considered as follows: (1) Creating auxiliary tasks can reduce the search difficulty; without prior knowledge, the difficulties faced by the subset of constraints included in the auxiliary tasks are different. In this case, an auxiliary task may not have a complementary evolutionary trial with the main task. In addition, when all constraints are eliminated, there is no infeasible region to hinder the evolution of the population. In this case, the created unconstrained MOP (i.e., a special auxiliary task) can help the main task cross the infeasible region through knowledge transfer.

[0046] (2) When generating multiple auxiliary tasks, optimizing them will consume more computing resources. Optimizing these auxiliary tasks requires more storage space and computing time. However, different from the general multi-task optimization problem, the computing resources are proportional to the number of tasks. No matter how many tasks the CMOP is decomposed into, the total amount of computing resources is fixed. When only one task is created, most of the computing resources can be used to solve the main task. In the simplified "special case" with only two tasks, from the perspective of the objective space, the main task is to find a well-distributed constrained Pareto front (CPF), while the auxiliary task is to find a well-distributed unconstrained Pareto front (UPF). Transferring content is an important issue for EMT because only transferring valuable knowledge is helpful for EMT. If the quality of the transferred knowledge is not distinguished, negative transfer will occur and the evolutionary direction will be disrupted. Therefore, it is necessary and important to find valuable knowledge.

[0047] In the population evolution process of the present invention, the knowledge of a task includes two types: parent individuals and offspring individuals. The parent individuals and offspring individuals respectively constitute the parent population and the offspring population. Different populations may be valuable knowledge at different evolutionary stages.

[0048] (1) The first stage: The parent population (ZQ1) of the set main task (T1) has a large infeasible region. Since the auxiliary task (denoted as T2) does not consider constraints, its parent population (denoted as ZQ2) is located in the infeasible region. Since the evolution of the population depends on better offspring individuals, in this stage, some parent individuals may be dominated by some offspring individuals. Therefore, ZQ1 and ZQ2 produce some offspring individuals with better objective values. However, ZQ1 is not easily able to escape from the local feasible region without knowledge transfer because its offspring individuals with better objective values are infeasible. In addition, when ZQ2 is transferred to T1, ZQ1 still cannot jump out of the local feasible region. On the contrary, the offspring individuals of ZQ2 are located in a better feasible region; therefore, when they are transferred to T1, ZQ1 can cross a large infeasible region. Therefore, in the first stage, the offspring population is marked as valuable knowledge.

[0049] Referring to Table 1, as shown in Algorithm 3.1, EMCMO starts with two populations ZQ1 and ZQ2 that focus on two tasks. Then, the two populations are evaluated on the corresponding tasks. During the main loop, since the useful knowledge is different in different stages, the evolutionary process is divided into two stages. In the first stage, the two populations evolve according to Algorithm 3.2. Then, in the second stage, they are updated through Algorithm 3.3. In the algorithm, the Constraint Dominance Principle (CDP) method and the Fast Non-Dominated Sorting method are used as the environmental selection methods for the two tasks.

[0050] 1) The first stage Referring to Table 2, Algorithm 3.2 describes the update process of EMCMO in the first stage. Among them, when one task is represented by T j (j = (1, 2)), the other task is represented by T 2 / j . In each generation, for each task T j , the parent for its crossover operation is randomly selected from the corresponding parental population P j to improve population diversity. High population diversity can prevent premature convergence. Then, NIP / 2 offspring individuals are generated using the crossed parents, where NIP is the size of the parental population. The offspring population is OP j . Combine ZQ j , OP j and OP 2 / j to obtain the temporary population TP j . Next, evaluate TP j for the corresponding task. Finally, through environmental selection, NIP individuals are selected from TP j to form a new population ZQ j . Among them, when evaluating an individual on T2, not only the objective function value but also the value of the constraint violation degree is obtained. In the first stage, population diversity is considered in the process of selecting crossed parents. Among them, the offspring population has higher diversity and better quality and is marked as valuable knowledge.

[0051] Table 1

[0052] Table 2

[0053] 2) The second stage: Compared with the first stage, the main work in the second stage is to determine the transferred population. Since it is difficult to know the relationship between CPF and UPF without prior knowledge, the present invention proposes a method to evaluate the effectiveness of the population. First, P with NIP individuals jCombine with OPj having NIP / 2 individuals to obtain a new population TOP with (3·NIP / 2) individuals j . Then, according to the environmental selection rule of T 2 / j , select the best NIP individuals in TOP j . Next, calculate the success rates of ZQ j and OP j on T 2 / j . The calculation formula is as follows:

[0054] Among them, ∈ [0, 1] and ∈ [0, 1] respectively represent the success rates of ZQ j and OP j ; num_ZQ j and numOP j are the number of parental individuals and the number of offspring individuals in the best NIP individuals respectively. When αZQ j < αOP j , it means that OP j is more suitable for T 2 / j . In this case, OP j is considered as the transfer population TrP j of T j . Otherwise, ZQ j will be a better choice for T 2 / j . In this case, in order to maintain the diversity of T 2 / j , the NIP / 2 individuals randomly selected from ZQ j are regarded as the transfer population TrP j of T j .

[0055] Referring to Table 3, Algorithm 3.3 gives the update process of EMCMO in the second stage. In each generation, for each task T j , select the parent to cross with it from ZQ j according to its quality to improve the convergence performance. These crossed parents are used to generate offspring individuals of NIP / 2. According to 3) and 4), the transfer population TrP j of T j is determined. Then, combine ZQ j , OPj and TrP 2 / j into a temporary population TP j . Finally, through environmental selection, select NIP individuals from TP j to obtain a new ZQ j .

[0056] Table 3

[0057] In order to ensure population diversity and improve the search ability, the present invention mutates some individuals staying on the boundary by the following method, and its expression is as follows:

[0058] Wherein 、 、 、 、 and i are randomly generated unequal integers; where F represents the mutation factor, a random number between 0 and 1 is assigned to F, and its main function is to control the degree of population mutation. The algorithm of the present invention proposes an improved boundary processing strategy, which mainly aims at some individuals staying on the boundary, and these individuals will be readjusted to be randomly distributed within the entire decision space range to improve the diversity of the distribution of population individuals during the search process. The following is this improved mutation strategy:

[0059] If the new individual generated after mutation using this improved mutation strategy still exceeds the boundary, the individual will be randomly distributed within the search space range.

[0060] The embodiment of the present invention provides a method for optimizing the industrial value chain under carbon constraints. First, a carbon-constrained industrial value chain is constructed, and then an optimization objective function of the carbon-constrained industrial value chain is constructed from the economic dimension, and constraint conditions of the carbon-constrained industrial value chain are constructed from the carbon emission dimension, so as to construct an industrial value chain model affected by carbon constraints. Then, a constrained multi-objective optimization algorithm based on evolutionary multi-task processing is used to solve the carbon-constrained industrial value chain model, and an optimal solution for constructing the industrial value chain under carbon constraints is obtained, so that the maximum benefit can be obtained under the condition of meeting carbon constraints.

[0061] The embodiment of the present invention also provides a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, the above method is implemented.

[0062] Those skilled in the art of this technology can understand that, unless otherwise defined, all terms (including technical terms and scientific terms) used here have the same meaning as the general understanding of those of ordinary skill in the art to which the present invention belongs. It should also be understood that those terms defined in a general dictionary, such as those, should be understood to have a meaning consistent with the meaning in the context of the prior art, and will not be interpreted with an idealized or overly formal meaning unless specifically defined.

[0063] For the method embodiments, for the sake of simplicity in description, they are all expressed as a series of combinations of actions. However, those skilled in the art should understand that the embodiments of the present invention are not limited by the described order of actions, because according to the embodiments of the present invention, some steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also understand that the embodiments described in the specification are all preferred embodiments, and the actions involved are not necessarily essential for the embodiments of the present invention.

[0064] From the description of the above embodiments, those skilled in the art can clearly understand that this application can be implemented by means of software plus a necessary general hardware platform. Based on such an understanding, the technical solution of this application, in essence, or the part that makes contributions to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a storage medium, such as ROM / RAM, magnetic disk, optical disc, etc., and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in various embodiments or some parts of the embodiments of this application.

[0065] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some or all of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for optimizing the industrial value chain under carbon constraints, characterized in that: include: Build a carbon-constrained industrial value chain; From the economic dimension, we construct the optimization objective function of the carbon-constrained industrial value chain, from the carbon emission dimension, we construct the constraint conditions of the carbon-constrained industrial value chain, and we construct the industrial value chain model under the influence of carbon constraints, where: （U） ； （L） ； Among them, (U) is the optimization objective equation group, (L) is the constraint objective equation group, f1 is the objective function of the optimization algorithm, x is the decision variable, and W is the constraint condition of the decision variable; f2 is the objective function in the constraint, y is the decision variable, and w is the constraint condition of the decision variable. The carbon emission constraint function is as follows: ； Among them, CZ represents the total carbon footprint of the product, CZ(i) represents the carbon footprint contribution of each enterprise, and CZ MAX represents the carbon footprint threshold of the product, and n is the number of enterprises; A constrained multi-objective optimization algorithm based on evolutionary multi-tasking is used to solve the carbon-constrained industrial value chain model. The main task and auxiliary task are set to transfer knowledge to improve the search ability of the population and thus improve the optimization ability of the algorithm. The constrained multi-objective optimization algorithm is used to solve the optimal solution for building the industrial value chain under carbon constraints, and finally the construction of the optimal industrial value chain is completed based on the optimal solution.

2. The method according to claim 1, characterized in that The construction of the carbon constraint industry value chain includes: G = {V, E} represents the industrial value chain, where V = {1, 2, ..., B} represents the set of node enterprises in the industrial value chain, and B represents the total number of nodes in the industrial value chain; E = {e1, ...e k } represents the set of cooperative relationships between different enterprises in the industrial value chain, including the following steps: At the initial state t=0, i.e. the time corresponding to the network state, A initial nodes are generated, and the edge connections between nodes are randomly generated; At time t, a new node is added, and the newly added candidate nodes of different categories and modules are distinguished according to the le value and lv value; among them, the le value is the category of the node, and the lv value is the module where the node is located, that is, the division of labor in the industrial value chain. The new edge connection rules are as follows: when le=1; lv=1, it means that the newly added node is a node of the first category and the first module, and it can only choose to establish a connection with the nodes of the second category and the first module; when le=2 and lv=2, it means that the newly added node is an enterprise node of the second category, and it can choose to establish a connection with the enterprise nodes of the first and third categories, and when connecting, it chooses to connect with the nodes of the second module in the first and third categories; At the tth moment, the newly added node selects the existing A in the network according to the new edge connection rule. t The probability of connecting a new node to an existing old node i for: ； It increases the degree of each node but does not affect the priority connection mechanism, where represents the degree of node i; represents the total degree of the network at time t; When the number of node data in the model reaches B, the number of new node data stops increasing.

3. The method according to claim 1, characterized in that The constraints for constructing the carbon-constrained industrial value chain from the carbon emission dimension include carbon emission intensity and processing costs: Among them, carbon emissions are: ； in, is carbon emissions; is the consumption of the a-th energy source; is the average lower heating value of the fuel; The carbon content per unit calorific value of the fuel; is the oxygen content; 44 / 12 is the conversion factor from C to CO2; The carbon emission intensity is: ； in, is the carbon emission intensity, Add value to the enterprise; The processing cost is: ； ； in, represents the processing cost of enterprise i; represents the cost of energy type j of enterprise i; Represents the innovation factor data of various energy sources, including coal, natural gas, non-fossil energy, and green energy; Indicates the carbon emissions corresponding to various energy sources; represents the price data of various energy sources of enterprise i; Represents the carbon reduction target of enterprise i.

4. The method according to claim 3, characterized in that The constraints for building a carbon-constrained industrial value chain from the carbon emission dimension also include: Obtain the overall Gini coefficient of the industry value chain, including: ； in, is the overall Gini coefficient, is the carbon emissions of the jth enterprise in the i-th module; is the carbon emission of the rth enterprise in the mth module; Z is the total number of modules; B is the total number of enterprises; is the number of enterprises in the ith module, is the number of enterprises in the mth module; is the average carbon emission of the enterprise; Obtain the intra-module Gini coefficient and inter-module Gini coefficient in the industrial value chain, including: ； in, is the Gini coefficient within the module, indicating the degree of carbon emission differences among enterprises within the module; is the inter-module Gini coefficient, which indicates the degree of difference in carbon emissions between different modules; is the average carbon emission of enterprises in module i, indicating the average carbon emission of enterprises in module i; is the average carbon emission value of enterprises in module m, indicating the average carbon emission value of enterprises in module m. is the carbon emissions of the rth enterprise in the ith module.

5. The method according to claim 4, characterized in that The constraints for building a carbon-constrained industrial value chain from the carbon emission dimension also include: Obtain the overall Gini coefficient of the industry value chain, including: ； in: , , They are the intra-module, inter-module and super-variable density difference contributions, indicating the difference contributions of carbon emissions in different dimensions; is the relative impact of carbon emissions between the ith and mth modules, indicating the relative impact between two specified modules; is the carbon emission difference between modules i and m, indicating the specific value of the carbon emission difference between the two specified modules; is the hypervariable first-order moment, is the proportion of carbon emissions of module i to the overall carbon emissions, is the proportion of carbon emissions of module m to the total carbon emissions.

6. The method according to claim 1, characterized in that The optimization objective function of building a carbon-constrained industry value chain from an economic perspective includes: The economic optimization goal is the profit function of the products in the industrial value chain, including the product price. , Procurement cost , Manufacturing Cost , the cost of value creation activities , and loss costs , industry value chain profit The calculation formula is as follows: ； ； ； ； ； in, The cost of purchasing raw materials for parts suppliers, the cost of procuring parts for manufacturers; The cost of the manufacturing process for raw material suppliers, The cost of manufacturing the machining process for the parts manufacturer; , , They are the value creation costs of raw material suppliers, parts suppliers, and manufacturers; , , They are the loss costs incurred by raw material suppliers, parts suppliers and manufacturers during their operations respectively.

7. The method according to claim 1, characterized in that Also includes: Create an unconstrained multi-objective optimization problem with zero constraints as a special auxiliary task.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method described in any one of claims 1 to 7 is implemented.

Images