Visual management system and method for boiler supply chain

Abstract

The invention relates to the technical field of supply chain management, in particular to a boiler supply chain visual management system and method, and the system comprises a data obtaining module which is used for obtaining the data of suppliers and purchasers; the order management module is used for predicting orders of purchasers; the warehouse management module is used for managing warehouse information; the transportation management module is used for collecting logistics data and optimizing routes between suppliers and warehouses; the supplier evaluation module is used for establishing a supplier comprehensive evaluation mechanism model; and the visual display module is used for visually displaying the data and providing an early warning function. According to the method, the order of the purchaser is predicted by using the acquired data of the suppliers and the purchaser, the route between the suppliers and the warehouse is optimized, then the supplier comprehensive evaluation mechanism model is established, and finally the data is visually presented, so that a comprehensive and accurate decision basis is provided for managers.

Description

Technical Field

[0001] The present invention relates to the technical field of supply chain management, and specifically, to a boiler supply chain visualization management system and method. Background Art

[0002] With the continuous growth of global energy demand, boilers, as important thermal energy equipment, are playing an increasingly important role in various industries; however, there are many problems in the traditional boiler supply chain management method, such as opaque information, low collaborative efficiency, resource waste, etc., which seriously restrict the sustainable development of the boiler industry.

[0003] In recent years, with the rapid development of information technology and Internet of Things technology, visualization management systems have been widely used in various fields; by displaying data in the form of graphs, charts, etc., visualization management systems can help users more intuitively understand the distribution and changes of data, thereby improving decision-making efficiency and accuracy; a boiler supply chain visualization management system refers to using information technology means to integrate, display, and analyze relevant data of the company's warehouse and supply chain to form intuitive visualization graphs and reports, and through transparent data-driven, maximize the efficient operation of the enterprise's logistics, value stream, data stream, and cash flow, and realize full-process digital management; establish a comprehensive evaluation mechanism model for suppliers, and carry out applications such as supplier health and evaluation based on the mechanism model, and conduct comprehensive performance analysis of supply chain management; such a system can enable managers to clearly understand data such as logistics processes, transportation conditions, and inventory situations, so as to better allocate resources and make decisions.

[0004] The existing supply chain visualization management systems only simply supervise, lacking measures such as predicting purchasers' orders and optimizing the routes between suppliers and warehouses, which provides limited help for managers' decision-making and is not conducive to managers formulating relevant business plans. Summary of the Invention

[0005] The purpose of the present invention is to provide a boiler supply chain visualization management system and method to solve the problems raised in the above background art.

[0006] To achieve the solution of the above technical problems, one of the purposes of the present invention is to provide a boiler supply chain visualization management system, including:

[0007] A data acquisition module: used to acquire various data of suppliers and purchasers;

[0008] An order management module: used to predict purchasers' orders;

[0009] A warehouse management module: used to manage warehouse information;

[0010] Transportation Management Module: Used to collect logistics data and optimize the routes between suppliers and warehouses;

[0011] Supplier Evaluation Module: Used to establish a comprehensive supplier evaluation mechanism model;

[0012] Visualization Display Module: Used to visually present data and provide a warning function.

[0013] As a further improvement of this technical solution, in the order management module, features with high correlation are removed based on Pearson correlation analysis. The corresponding steps are as follows:

[0014] Construct a feature vector using the historical data of purchasers obtained by the data acquisition module. There are m pieces of data, and each piece of data has n0 features. Among them, the kth normalized data is expressed as

[0015] For any two features x and y among the n0 features in the data, calculate the correlation coefficient r xy , then

[0016]

[0017] Among them, and respectively represent the average values of a k,x and a k,y . Set a threshold When , retain one of the features. Each piece of data obtained has n1 features, and the kth data is expressed as

[0018] As a further improvement of this technical solution, the order management module predicts orders based on an improved random forest algorithm. Divide the m pieces of data into a training set and a validation set. Use Boostrap to extract c samples from the training set, randomly extract n features to construct a decision tree, select the optimal splitting attribute according to the Gini index, and continue to split downward until reaching the depth D of the decision tree or all samples at this node are of the same category. Construct N decision trees. The calculation formula for the Gini index is:

[0019]

[0020] Among them, G s represents the Gini coefficient of node s on the decision tree, R represents the total number of categories, and p rs represents the proportion of samples belonging to the rth category on node s of the decision tree.

[0021] As a further improvement of this technical solution, in the order management module, the number of decision trees N, the number of extracted features n, and the depth D of the decision tree in the random forest are optimized based on the firefly algorithm. The specific steps are as follows:

[0022] Population initialization. Suppose there are h firefly individuals, and the position of individual p is denoted as h p =(N p ,n p ,D p ). Its brightness is the objective function value f(h p ) corresponding to this position. The random forest is trained using the parameters corresponding to the firefly individuals in the population, and the accuracy rate is used as the fitness function. The relative brightness I between individuals p and q is:

[0023] I = I0exp(-γr p,q )

[0024]

[0025] where the position of individual q is denoted as h q =(N q ,n q ,D q ), I0 is the initial firefly brightness, γ is the light absorption ability coefficient, and r p,q is the spatial distance between two fireflies;

[0026] The attractiveness β between individuals p and q is:

[0027] β = β0exp(-γr p,q )

[0028] where β0 is the initial attractiveness;

[0029] The position update formula for individual p to move towards q is:

[0030] h p = h q +β(h p -h q )+α(rand - 0.5)

[0031] where α is the step factor and rand is a random factor uniformly distributed on [0, 1];

[0032] Adaptive light absorption ability coefficient γ:

[0033]

[0034] where γ0 is the initial light absorption ability coefficient, g is the current iteration number, and G is the maximum iteration number.

[0035] As a further improvement of this technical solution, the transportation management module optimizes the route between suppliers and warehouses based on an improved genetic algorithm. The specific steps are as follows:

[0036] Use the data acquisition module to obtain supplier information and warehouse information;

[0037] Initialize the population: Set the population size, use real number coding, and the gene sequence number of the individual chromosome represents the moving path of the truck. Suppose there are l a suppliers and l b warehouses. The chromosome length is equal to (l a +l b ). Each gene corresponds to the supplier or warehouse number, and the warehouses are evenly distributed among the suppliers;

[0038] Fitness evaluation: Use the reciprocal of the transportation cost as the fitness function value;

[0039] Selection: Adopt tournament selection and roulette wheel selection, with the proportions being θ and (1 - θ) respectively;

[0040] Crossover: Select two individuals for crossover operation, using the adaptive crossover probability P c , then

[0041]

[0042] where k1 and k2 are set probabilities, 0 < k1, k2 < 1, F max is the maximum fitness value of the individuals in the population, F avg is the average fitness value of the population, and F1 is the larger fitness value of the two crossover individuals;

[0043] Mutation: Select an individual for mutation operation, using the adaptive mutation probability P m , then

[0044]

[0045] where k3 and k4 are set probabilities, 0 < k3, k4 < 1, and F2 is the fitness value of the mutated individual;

[0046] Population update: Update the fitness values of the individuals in the sub-population;

[0047] Termination condition judgment: Stop the algorithm when the maximum number of iterations is reached, and output the optimal result.

[0048] As a further improvement of this technical solution, the supplier evaluation module establishes a comprehensive supplier evaluation mechanism model based on the improved TOPSIS model. The specific steps are as follows:

[0049] Suppose there are u suppliers to be evaluated, with v evaluation indicators, and the j-th indicator value of the i-th supplier is t ij , the initial matrix T is established as follows:

[0050]

[0051] Positive normalization of indicators:

[0052] For extremely large indicators:

[0053] o ij = t ij

[0054] For extremely small indicators:

[0055] o ij = t ij (max) - t ij

[0056] where t ij (max) represents the maximum value of this column;

[0057] Normalization processing:

[0058]

[0059] The normalized matrix W is:

[0060]

[0061] The weighted decision matrix S is:

[0062]

[0063] where ε1, ε2, …, ε v are the comprehensive weights of each evaluation indicator respectively; calculate the positive ideal solution S + :

[0064]

[0065] where, are the maximum values of each column respectively;

[0066] Calculate the negative ideal solution S - :

[0067]

[0068] where, are the minimum values of each column respectively;

[0069] Calculate the Euclidean distance:

[0070]

[0071] Among them, and respectively represent the distances from the positive ideal solution and the negative ideal solution;

[0072] Calculate the TOPSIS evaluation value z i of the i-th supplier, then

[0073]

[0074] As a further improvement of this technical solution, the supplier evaluation module establishes a comprehensive supplier evaluation mechanism model based on the improved TOPSIS model, and the comprehensive weights of each evaluation index are assigned using a combination of subjective weights and objective weights. Then

[0075]

[0076] where ε j represents the comprehensive weight of the j-th evaluation index, and ε j represents the comprehensive weight of the j-th evaluation index, σ j represents the subjective weight of the j-th evaluation index, and τ j represents the objective weight of the j-th evaluation index.

[0077] As a further improvement of this technical solution, the analytic hierarchy process is used in the supplier evaluation module to obtain the subjective weight σ j , and the specific steps are as follows:

[0078] Compare the importance degrees of each evaluation index, construct the judgment matrix H, and quantify the comparison results using the 1-9 scale method. Then

[0079]

[0080] where h 11 =h 22 =...=h vv =1, and h 1v represents the importance degree of the first evaluation index relative to the v-th evaluation index;

[0081] Hierarchical single sorting and consistency test, calculate the consistency index CI. Then

[0082]

[0083] where λ max is the maximum eigenvalue of the judgment matrix H;

[0084] Calculate the consistency ratio CR. Then

[0085]

[0086] Among them, RI is the random consistency index. When the consistency ratio CR is less than 0.1, the consistency test passes; otherwise, corrections are made.

[0087] For weight calculation, the eigenvector method is adopted:

[0088] Hτ = λ max τ

[0089] Among them, τ is the eigenvector corresponding to the judgment matrix H. By normalizing the eigenvector τ, the subjective weights τ1, τ2, …, τ of each evaluation index are obtained. v .

[0090] As a further improvement of this technical solution, the entropy weight method is used in the supplier evaluation module to obtain the objective weight. The specific steps are as follows:

[0091] Calculate the proportion p of the i-th supplier under the j-th evaluation index in this evaluation index ij , then

[0092]

[0093] Calculate the entropy value Ej of the j-th evaluation index j , then

[0094]

[0095] Calculate the difference coefficient Gj of the j-th evaluation index j , then

[0096] G j = 1 - Ej j

[0097] Calculate the objective weight τj of the j-th evaluation index j , then

[0098]

[0099] The second object of the present invention is to provide a boiler supply chain visualization management system. According to the above boiler supply chain visualization management system, it includes the following steps:

[0100] S1. Obtain various data of suppliers and purchasers;

[0101] S2. Based on the improved random forest, predict the orders of purchasers after removing highly correlated features;

[0102] S3. Collect logistics data and optimize the route between suppliers and warehouses based on the improved genetic algorithm;

[0103] S4. Establish a comprehensive supplier evaluation mechanism model based on the improved TOPSIS model;

[0104] S5. Visualize and present data and provide a warning function.

[0105] Compared with the prior art, the beneficial effects of the present invention are as follows: In the boiler supply chain visualization management system and method of the present invention, various data of suppliers and purchasers are used to predict the orders of purchasers, optimize the route between suppliers and warehouses, establish a comprehensive supplier evaluation mechanism model, and finally visualize and present the data, providing comprehensive and accurate decision-making basis for managers. Brief Description of the Drawings

[0106] Figure 1 It is the system structure diagram of the present invention;

[0107] Figure 2 It is the method flow chart of the present invention.

[0108] The meanings of the various marks in the figure are as follows: 100, data acquisition module; 200, order management module; 300, warehouse management module; 400, transportation management module; 500, supplier evaluation module; 600, visualization display module. Detailed Embodiments

[0109] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying 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 the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without making creative efforts shall fall within the protection scope of the present invention.

[0110] As Figure 1 shown, this embodiment provides a boiler supply chain visualization management system, including:

[0111] Data acquisition module 100: used to acquire various data of suppliers and purchasers, and the purchasers are mainly warehouse data;

[0112] Order management module 200: used to predict the orders of purchasers to facilitate management and coordination by managers;

[0113] Warehouse management module 300: used to manage warehouse information, display information such as the inventory situation and storage conditions of the warehouse, and managers can adjust the warehousing strategy in a timely manner to improve warehousing efficiency;

[0114] Transportation management module 400: used to collect logistics data and optimize the route between suppliers and warehouses, and track the positions and statuses of vehicles and goods in real time. Managers can understand the transportation process of goods at any time and solve possible problems in a timely manner;

[0115] Supplier Evaluation Module 500: Used to establish a comprehensive supplier evaluation mechanism model to provide managers with comprehensive and accurate decision-making basis;

[0116] Visualization Display Module 600: Used to visually present data and provide a warning function, setting various warning indicators. When the indicators reach the warning line, the system will automatically issue a warning to remind managers to take measures in a timely manner.

[0117] In this embodiment, the steps corresponding to removing highly correlated features based on Pearson correlation analysis in the order management module 200 are as follows:

[0118] Construct a feature vector using the historical data of purchasers obtained by the data acquisition module 100. There are m pieces of data, and each piece of data has n0 features. Among them, the kth piece of data after normalization is expressed as

[0119] For any two features x and y among the n0 features in the data, calculate the correlation coefficient r xy , then

[0120]

[0121] Among them, and respectively represent the average values of a k,x and a k,y . Set the threshold When , retain one of the features. Each piece of data obtained has n1 features, and the kth piece of data is expressed as The Person correlation coefficient represents the degree of linear relationship between two variables. The closer the Person correlation coefficient is to 0, the less linear relationship there is between the two variables. Removing features with high linear correlation can avoid data complexity and reduce the calculation amount.

[0122] Predict the order based on the improved random forest algorithm. Divide the m pieces of data into a training set and a validation set. Use Boostrap to extract c samples from the training set, randomly extract n features to construct a decision tree, select the optimal splitting attribute according to the Gini index, and continue to split downwards until reaching the depth D of the decision tree or all samples at this node are of the same category. Construct N decision trees. The calculation formula of the Gini index is:

[0123]

[0124] Among them, G s represents the Gini coefficient of node s on the decision tree, R represents the total number of categories, and p rs represents the proportion of samples belonging to the rth category on node s of the decision tree.

[0125] Optimize the number of decision trees \(N\), the number of extracted features \(n\), and the decision tree depth \(D\) in the random forest based on the firefly algorithm. These three parameters have the greatest impact on the regression results of the random forest. Using the firefly algorithm can improve the accuracy of the prediction results of the random forest. The specific steps are as follows:

[0126] Population initialization. Suppose there are \(h\) firefly individuals, and the position of individual \(p\) is denoted as \(h\) p =(N p ,n p ,D p ). Its brightness is the objective function value \(f(h\) p ) corresponding to this position. Train the random forest using the parameters corresponding to the firefly individuals in the population, and use the accuracy rate as the fitness function. The relative brightness \(I\) between individuals \(p\) and \(q\) is:

[0127] I = I0exp(-γr p,q )

[0128]

[0129] Among them, the position of individual \(q\) is denoted as \(h\) q =(N q ,n q ,D q ). I0 is the initial firefly brightness, γ is the light absorption ability coefficient, and \(r\) p,q is the spatial distance between two fireflies;

[0130] The attraction degree β between individuals \(p\) and \(q\) is:

[0131] β = β0exp(-γr p,q )

[0132] Among them, β0 is the initial attraction degree;

[0133] The position update formula for individual \(p\) to move towards \(q\) is:

[0134] h p =h q +β(h p -h q )+α(rand - 0.5)

[0135] Among them, α is the step size factor, and rand is a random factor uniformly distributed on [0,1];

[0136] Adaptive light absorption ability coefficient γ:

[0137]

[0138] Among them, γ0 is the initial light absorption ability coefficient, g is the current iteration number, G is the maximum iteration number, and the self-adaptive light absorption ability coefficient improvement algorithm converges the convergence ability.

[0139] In this embodiment, the transportation management module 400 optimizes the route between the supplier and the warehouse based on the improved genetic algorithm. The specific steps are as follows:

[0140] Use the data acquisition module 100 to obtain supplier information and warehouse information;

[0141] Initialize the population: Set the population size, use real number coding, and the individual chromosome gene number represents the truck movement path. Suppose there are l a suppliers and l b warehouses. The chromosome length is equal to (l a +l b ). Each gene corresponds to the supplier or warehouse number. The warehouses are evenly distributed among the suppliers. If they cannot be evenly distributed, then try to make the number of suppliers before and after the warehouse as close as possible. If there are 4 suppliers and 2 warehouses, the supplier numbers are 1 to 4, and the warehouse numbers are 5 and 6. One chromosome can be 1, 3, 6, 2, 4, 5;

[0142] Fitness evaluation: Use the reciprocal of the transportation cost as the fitness function value;

[0143] Selection: Adopt tournament selection and roulette wheel selection, and the proportions are θ and (1 - θ) respectively;

[0144] Crossover: Select two individuals for crossover operation, use the self-adaptive crossover probability P c , then

[0145]

[0146] Among them, k1 and k2 are set probabilities, 0 < k1, k2 < 1, F max is the maximum fitness value of the individuals in the population, F avg is the average fitness value of the population, and F1 is the larger fitness value of the two crossover individuals, which improves the optimization ability of the algorithm;

[0147] Mutation: Select one individual for mutation operation, use the self-adaptive mutation probability P m , then

[0148]

[0149] Among them, k3 and k4 are set probabilities, 0 < k3, k4 < 1, and F2 is the fitness value of the mutated individual, which improves the optimization ability of the algorithm;

[0150] Population update: Update the fitness values of the individuals in the sub-population;

[0151] Termination condition judgment: Stop the algorithm when the maximum number of iterations is reached, and output the optimal result.

[0152] In this embodiment, the supplier evaluation module 500 establishes a comprehensive supplier evaluation mechanism model based on the improved TOPSIS model. The specific steps are as follows:

[0153] Suppose there are u suppliers to be evaluated and v evaluation indicators. Common evaluation indicators include product quality, product cost, delivery ability, after-sales service, technical ability, and financial status, etc. The value of the jth indicator of the ith supplier is t ij , Using the objective data of the survey, establish the initial matrix T as:

[0154]

[0155] Index positive normalization processing:

[0156] For extremely large indicators:

[0157] o ij = t ij

[0158] For extremely small indicators:

[0159] o ij = t ij (max) - t ij

[0160] Among them, t ij (max) represents the maximum value of this column;

[0161] Normalization processing:

[0162]

[0163] The normalized matrix W is:

[0164]

[0165] The weighted decision matrix S is:

[0166]

[0167] Among them, ε1, ε2, …, ε v are the comprehensive weights of each evaluation indicator respectively;

[0168] Calculate the positive ideal solution S + :

[0169]

[0170] Among them, They are the maximum values of each column respectively;

[0171] Calculate the negative ideal solution S - :

[0172]

[0173] where, They are the minimum values of each column respectively;

[0174] Calculate the Euclidean distance:

[0175]

[0176] where, and represent the distances from the positive ideal solution and the negative ideal solution respectively;

[0177] Calculate the TOPSIS evaluation value z of the i-th supplier i , then

[0178]

[0179] The supplier evaluation module 500 establishes a comprehensive supplier evaluation mechanism model based on the improved TOPSIS model. The comprehensive weights of each evaluation index are assigned by combining subjective weights and objective weights. The comprehensive weights comprehensively consider subjective weights and objective weights, can fully reflect the importance of evaluation indexes and the information volume of actual data, making the evaluation results more comprehensive. Subjective weights are often affected by the subjective experience and cognition of evaluators, which may lead to unreasonable weight allocation, while objective weights are calculated based on actual data, avoiding the influence of subjective factors. By combining subjective weights and objective weights, the influence of subjective factors can be reduced to a certain extent and the objectivity of evaluation can be improved. Then

[0180]

[0181] where, σ j represents the subjective weight of the j-th evaluation index, and τ j represents the objective weight of the j-th evaluation index.

[0182] Use the analytic hierarchy process to obtain the subjective weight σ j , and the specific steps are as follows:

[0183] Compare the importance degrees of each evaluation index, make a questionnaire survey, distribute it to experts and managers, and quantify the comparison results according to the constructed judgment matrix H using the 1-9 scale method. Then

[0184]

[0185] where, h11 = h 22 = … = h vv = 1, h 1v represents the importance degree of the first evaluation index relative to the v-th evaluation index;

[0186] Hierarchical single sorting and consistency test. Calculate the consistency index CI, then

[0187]

[0188] where λ max is the maximum eigenvalue of the judgment matrix H;

[0189] Calculate the consistency ratio CR, then

[0190]

[0191] where RI is the random consistency index. When the consistency ratio CR is less than 0.1, the consistency test passes; otherwise, corrections are made;

[0192] Weight calculation. The eigenvector method is adopted:

[0193] Hτ = λ max τ

[0194] where τ is the eigenvector corresponding to the judgment matrix H. Normalize the eigenvector τ to obtain the subjective weights τ1, τ2, …, τ v .

[0195] Use the entropy weight method to obtain the objective weights. The specific steps are as follows:

[0196] Calculate the proportion p of the i-th supplier under the j-th evaluation index in this evaluation index ij , then

[0197]

[0198] Calculate the entropy value E of the j-th evaluation index j , when p ij = 0, p ij ln p ij = 0, then

[0199]

[0200] Calculate the difference coefficient Gj of the j-th evaluation index j , then

[0201] G j = 1 - E j

[0202] Calculate the objective weight τ of the j-th evaluation index j , then

[0203]

[0204] As Figure 2 shown, this embodiment also provides a visual management method for the boiler supply chain. According to the above-mentioned boiler supply chain visualization management system, it includes the following steps:

[0205] S1. Obtain various data of suppliers and purchasers;

[0206] S2. Remove highly correlated features and then predict the orders of purchasers based on the improved random forest;

[0207] S3. Collect logistics data and optimize the route between suppliers and warehouses based on the improved genetic algorithm;

[0208] S4. Establish a comprehensive evaluation mechanism model for suppliers based on the improved TOPSIS model;

[0209] S5. Visually present the data and provide a warning function.

[0210] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. The above embodiments and descriptions in the specification are only preferred examples of the present invention and are not used to limit the present invention. Without departing from the spirit and scope of the present invention, the present invention will have various changes and improvements, and these changes and improvements all fall within the scope of the present invention claimed. The scope of protection claimed by the present invention is defined by the appended claims and their equivalents.

Claims

1. A visual management system for a boiler supply chain, characterized in that, Including: Data acquisition module (100): used to acquire various data of suppliers and purchasers; Order management module (200): used to predict the orders of purchasers; Warehouse management module (300): used to manage warehouse information; Transportation management module (400): used to collect logistics data and optimize the route between suppliers and warehouses; Supplier evaluation module (500): used to establish a comprehensive supplier evaluation mechanism model; Visualization display module (600): used to visually present data and provide a warning function.

2. The boiler supply chain visualization management system according to claim 1, wherein: In the order management module (200), features with high correlation are removed based on Pearson correlation analysis. The corresponding steps are as follows: Construct a feature vector using the historical data of purchasers obtained by the data acquisition module (100). There are m pieces of data, and each piece of data has n0 features. The k-th piece of data after normalization is expressed as For any two features x and y among the n0 features in the data, calculate the correlation coefficient r xy , then Among them, and respectively represent the average value of a k,x and a k,y . Set a threshold When , retain one of the features. Each obtained data has n1 features, and the k-th data is represented as 3. The boiler supply chain visualization management system according to claim 1, characterized in that: The order management module (200) predicts orders based on an improved random forest algorithm. It divides m data into a training set and a validation set, extracts c samples from the training set using Bootstrap, randomly selects n features to construct a decision tree, selects the optimal splitting attribute according to the Gini index, and continues to split downwards until the depth D of the decision tree is reached or all samples of this node are of the same category. N decision trees are constructed. The calculation formula of the Gini index is: Among them, G s represents the Gini coefficient of node s on the decision tree, R represents the total number of categories, and p rs represents the proportion of samples belonging to the r-th category on node s of the decision tree.

4. The boiler supply chain visualization management system according to claim 1, characterized in that: In the order management module (200), the number of decision trees N, the number of extracted features n, and the depth D of the decision tree in the random forest are optimized based on the firefly algorithm. The specific steps are as follows: Population initialization. Suppose there are h firefly individuals, and the position of individual p is denoted as h p =(N p ,n p ,D p ). Its brightness is the objective function value f(h p ) corresponding to this position. Use the parameters corresponding to the firefly individuals in the population to train the random forest, and take the accuracy as the fitness function. The relative brightness I between individuals p and q is: I = I0 exp(-γr p,q ) Among them, the position of individual q is denoted as h q =(N q , n q , D q ), I0 is the initial firefly brightness, γ is the light absorption ability coefficient, and r p,q is the spatial distance between two fireflies; The attraction degree β between individuals p and q is: β = β0exp(-γr p,q ) Among them, β0 is the initial attraction degree; The position update formula for individual p to move towards q is: h p = h q + β(h p - h q ) + α(rand - 0.5) Among them, α is the step size factor, and rand is a random factor uniformly distributed on [0, 1]; Adaptive light absorption ability coefficient γ: Among them, γ0 is the initial light absorption ability coefficient, g is the current iteration number, and G is the maximum iteration number.

5. The boiler supply chain visualization management system according to claim 1, wherein: The transportation management module (400) optimizes the route between suppliers and warehouses based on an improved genetic algorithm. The corresponding specific steps are as follows: Use the data acquisition module (100) to acquire supplier information and warehouse information; Initialize the population: Set the population size, use real number coding, and the individual chromosome gene sequence number represents the moving path of the truck. There are l a suppliers and l b warehouses. The chromosome length is equal to (l a +l b ). Each gene corresponds to the sequence number of a supplier or a warehouse, and the warehouses are evenly distributed among the suppliers; Fitness evaluation: Use the reciprocal of the transportation cost as the fitness function value; Selection: Adopt tournament selection and roulette wheel selection, with the proportions being θ and (1 - θ) respectively; Crossover: Select two individuals for the crossover operation and use the adaptive crossover probability P c , then Among them, k1 and k2 are set probabilities, 0 < k1, k2 < 1, F max is the maximum fitness value of individuals in the population, F avg is the average fitness value of the population, and F1 is the larger fitness value among the two crossed individuals; Mutation: Select an individual for mutation operation and use the adaptive mutation probability P m , then Among them, k3 and k4 are set probabilities, 0 < k3, k4 < 1, and F2 is the fitness value of the mutated individual; Population update: Update the fitness values of individuals in the sub-population; Termination condition judgment: Stop the algorithm when the maximum iteration number is reached, and output the optimal result.

6. The boiler supply chain visualization management system according to claim 1, wherein: The supplier evaluation module (500) establishes a comprehensive supplier evaluation mechanism model based on an improved TOPSIS model. The corresponding specific steps are as follows: Suppose there are \(u\) suppliers to be evaluated, with \(v\) evaluation indicators. The value of the \(j\)th indicator of the \(i\)th supplier is \(t\). ij , and the initial matrix \(T\) is established as follows: Index positive transformation: For extremely large type indicators: o ij = t ij For extremely small type indicators: o ij = t ij (max) - t ij where t ij (max) represents the maximum value of this column; Standardization processing: The normalized matrix W is: The weighted decision matrix S is: Among them, ε1, ε2, …, ε v are the comprehensive weights of each evaluation index respectively; calculate the positive ideal solution S + : wherein, are respectively the maximum values of each column; Calculate the negative ideal solution S - : wherein, are the minimum values of each column respectively; Calculate the Euclidean distance: Among them, and respectively represent the distances to the positive ideal solution and the negative ideal solution; Calculate the TOPSIS evaluation value \(z\) of the \(i\)-th supplier i , then 7. The boiler supply chain visualization management system according to claim 1, wherein: The supplier evaluation module (500) establishes a comprehensive supplier evaluation mechanism model based on an improved TOPSIS model. The comprehensive weights of each evaluation index are assigned using a combination of subjective weights and objective weights. Then Among them, ε j represents the comprehensive weight of the j-th evaluation index, and σ j represents the subjective weight of the j-th evaluation index, and τ j represents the objective weight of the j-th evaluation index.

8. The boiler supply chain visualization management system according to claim 7, wherein: The subjective weight σ is obtained by using the analytic hierarchy process in the supplier evaluation module (500). j , and the specific steps are as follows: Compare the importance degrees of each evaluation index, construct a judgment matrix H, and quantify the comparison results using the 1 - 9 scale method. Then where h 11 = h 22 =... = h vv = 1, h 1v represents the importance degree of the first evaluation index relative to the v-th evaluation index; Hierarchical single sorting and consistency test, calculate the consistency index CI. Then Among them, λ max is the maximum eigenvalue of the judgment matrix H; Calculate the consistency ratio CR. Then Among them, RI is the random consistency index. When the consistency ratio CR is less than 0.1, the consistency test passes; otherwise, corrections are made. Weight calculation is performed using the eigenvector method: Hτ = λ max τ Among them, τ is the eigenvector corresponding to the judgment matrix H. By normalizing the eigenvector τ, the subjective weights τ1, τ2, …, τ of each evaluation index can be obtained. v .

9. The boiler supply chain visualization management system according to claim 7, characterized in that: In the supplier evaluation module (500), the entropy weight method is used to obtain the objective weight. The specific steps are as follows: Calculate the proportion p of the i-th supplier in the j-th evaluation index ij , then Calculate the entropy value \(E\) of the \(j\)th evaluation index j , then Calculate the coefficient of variation G of the j-th evaluation index j , then G j = 1 - E j Calculate the objective weight τ of the j-th evaluation index j , then 10. A visual management method for a boiler supply chain, according to the boiler supply chain visual management system described in any one of claims 1-9, characterized in that, It includes the following steps: S1. Obtain various data of suppliers and purchasers. S2. After removing highly correlated features, predict the purchaser's orders based on the improved random forest. S3. Collect logistics data and optimize the route between the supplier and the warehouse based on the improved genetic algorithm. S4. Establish a comprehensive supplier evaluation mechanism model based on the improved TOPSIS model. S5. Visualize the data and provide a warning function.

Images