Bulk cargo international freight intelligent grouping optimization system based on data processing

Abstract

The invention provides a bulk cargo international freight intelligent grouping optimization system based on data processing, and relates to the technical field of data processing, and the system comprises an intelligent analysis module which is used for carrying out the analysis of a standardized data set through a deep neural network, so as to recognize a potential grouping opportunity and an optimization scheme; the grouping optimization module is used for determining a preliminary grouping combination based on the potential grouping opportunities and optimization schemes; for each preliminary group combination, adjusting the cost saving and transportation aging prediction of the preliminary group combination by using a dynamic adjustment factor to obtain an optimized group combination; and generating a final group splicing suggestion based on the optimized group splicing combination. According to the invention, the efficiency is improved, the cost is reduced, and the flexibility is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to an intelligent grouping optimization system for bulk international freight based on data processing. Background Art

[0002] In the field of traditional bulk international freight, group transportation of goods is an important link, which can effectively improve transportation efficiency and reduce transportation costs.

[0003] However, some traditional solutions only consider cost factors when optimizing group purchases, while ignoring other important factors such as transportation time, space utilization, etc. This may result in reduced transportation costs but longer transportation time, or low space utilization, thus affecting overall logistics efficiency.

[0004] Furthermore, some traditional solutions lack flexibility and real-time performance. Once the group purchase plan is determined, it is difficult to dynamically adjust it based on real-time supply information, transportation demand, shipping schedule information, etc. This is obviously a significant flaw in the rapidly changing international freight environment. Summary of the invention

[0005] The technical problem to be solved by the present invention is to provide an intelligent grouping optimization system for international bulk freight based on data processing, which improves efficiency, reduces costs and increases flexibility.

[0006] In order to solve the above technical problems, the technical solution of the present invention is as follows: In the first aspect, a bulk cargo international freight intelligent grouping optimization system based on data processing includes: Data collection module, used to collect real-time data related to bulk international freight, including cargo source information, transportation demand, ship schedule information, space information and cost information; A data processing module for preprocessing data related to bulk international freight to form a standardized data set; Intelligent analysis module, which uses deep neural networks to analyze standardized data sets to identify potential group buying opportunities and optimization solutions; The group purchase optimization module is used to determine preliminary group purchase combinations based on potential group purchase opportunities and optimization solutions; for each preliminary group purchase combination, the dynamic adjustment factor is used to adjust its cost savings and transportation time prediction to obtain an optimized group purchase combination; based on the optimized group purchase combination, the final group purchase suggestion is generated; The user interaction module is used to display group buying suggestions to users and receive user feedback and instructions to adjust and optimize group buying plans; The execution module is used to automatically execute the group buying operation according to the group buying plan confirmed by the user, including space reservation, cargo transportation arrangement and fee settlement.

[0007] Furthermore, a deep neural network is used to analyze the standardized data set to identify potential group buying opportunities and optimization solutions, including: Determine the structure of the deep neural network model, including determining the configuration of the input layer, hidden layer, and output layer; Divide the standardized data set into training set, validation set and test set, and use the training set to train the deep neural network model; adjust the weight of the deep neural network model through the gradient descent algorithm; evaluate the performance of the deep neural network on the validation set, monitor the overfitting during the training process, and adjust the hyperparameters as needed; evaluate the performance of the trained deep neural network model on the test set to obtain the evaluation results; optimize the deep neural network model according to the evaluation results, and repeat the training and validation process to obtain the optimized deep neural network model; Use the optimized deep neural network model to predict new standardized data to identify potential group buying opportunities; analyze the cost savings and efficiency improvement potential between different source combinations based on potential group buying opportunities, and screen practical and feasible group buying opportunities; Generate specific group buying optimization plans based on actual feasible group buying opportunities.

[0008] Furthermore, the weights of the deep neural network model are adjusted by the gradient descent algorithm, including: Randomly initialize the weight matrix and bias vector for each layer of the deep neural network model; Select a training sample as input and forward propagate it through the deep neural network model; At each layer, the dot product of the input and the weight is calculated, and the bias is added. Then the corresponding output data is obtained through the activation function, and the output data is used as the input of the next layer, and so on, until the prediction result of the final output layer is obtained; Use mean squared error to calculate the difference between the deep neural network model's prediction and the true value; The gradient of the output layer with respect to the weight is calculated according to the mean square error. Using the chain rule, the gradient of each layer with respect to the weight is calculated back layer by layer starting from the output layer. The weight and bias of each layer are updated according to the calculated gradient and the set learning rate.

[0009] Furthermore, based on potential group buying opportunities and optimization plans, determine the initial group buying combination, including: Set the number of individuals in the population, each individual represents a group combination, and the group combination is a list containing different source identifiers; randomly generate the list so that the sources in each list meet the basic compatibility requirements until the preset conditions are met to form the initial population; Construct a fitness function, apply the fitness function to each individual in the initial population, and calculate the fitness value of each individual; According to the fitness value of the individual, a roulette wheel selection method is used to select some individuals from the population as parents, the selected parent individuals are paired, and a single-point crossover is used to generate new offspring individuals; the generated offspring individuals are randomly mutated, and the parents, offspring and mutated individuals are combined into a new population, and the crossover and mutation operations are repeated until the termination condition is met, the iteration is stopped, and the corresponding individuals in the current population are output as the final solution; Convert the final solution into the actual puzzle combination.

[0010] Furthermore, the calculation formula of the dynamic adjustment factor is: ; in, Indicates The cost of a group combination; Indicates The cost of a group purchase combination in the baseline case; Indicates the number of group combinations; Indicates the transportation time of the group purchase combination; Indicates the transportation time of the benchmark group purchase combination; An adjustment factor that represents the effect of time on cost; Indicates The shipping time of each group purchase combination; Indicates The shipping time of each group purchase combination under the benchmark situation; Indicates the coefficient of impact of cost on transportation time; and is the weight coefficient.

[0011] Furthermore, for each preliminary group purchase combination, the dynamic adjustment factor is used to adjust its cost savings and transportation time prediction to obtain an optimized group purchase combination, including: For each preliminary group purchase combination, calculate its original cost savings forecast value, that is, the difference between the cost of the group purchase combination and the benchmark cost; Calculate the original transportation time forecast of the preliminary group combination, that is, the difference between the estimated transportation time of the group combination and the benchmark transportation time; Multiply the original cost savings forecast by the dynamic adjustment factor to obtain the adjusted cost savings forecast; Multiply the original transportation time prediction by the dynamic adjustment factor to obtain the adjusted transportation time prediction value; Combine the adjusted cost savings forecast and the adjusted shipping time forecast to form a comprehensive forecast result; Based on the comprehensive prediction results, each preliminary group purchase combination is comprehensively evaluated to obtain the evaluation results; According to the evaluation results, the group buying combination is optimized to obtain the optimized group buying combination.

[0012] Furthermore, based on the optimized group purchase combination, the final group purchase suggestion is generated, including: For each optimized group purchase combination, the weighted score is obtained by multiplying the adjusted cost savings forecast value and the adjusted transportation time forecast value by their respective weights and adding them together; The weighted scores are used as the sorting basis to sort the optimized group purchase combinations from high to low to obtain a sorted group purchase combination list; In the sorted group combination list, filter according to the preset filtering conditions; From the screened group purchase combinations, the corresponding combination is selected as the final group purchase suggestion based on the weighted score.

[0013] In the second aspect, a method for optimizing intelligent grouping of bulk international freight based on data processing includes: Real-time collection of data related to bulk international freight, including cargo source information, transportation demand, ship schedule information, space information and cost information; Preprocessing of data related to bulk international shipping to form a standardized data set; Use deep neural networks to analyze standardized data sets to identify potential group buying opportunities and optimization solutions; Based on potential group buying opportunities and optimization plans, determine preliminary group buying combinations; for each preliminary group buying combination, use dynamic adjustment factors to adjust its cost savings and transportation time prediction to obtain an optimized group buying combination; based on the optimized group buying combination, generate the final group buying suggestion; Show group buying suggestions to users and receive feedback and instructions from users to adjust and optimize group buying plans; According to the group buying plan confirmed by the user, the group buying operation is automatically executed, including space reservation, cargo transportation arrangement and fee settlement.

[0014] According to a third aspect, a computing device includes: one or more processors; The storage device is used to store one or more programs. When the one or more programs are executed by the one or more processors, the one or more processors implement the system.

[0015] In a fourth aspect, a computer-readable storage medium stores a program, and the program implements the system when executed by a processor.

[0016] The above scheme of the present invention includes at least the following beneficial effects.

[0017] Through the data collection module, the system can collect all kinds of data related to bulk international freight in real time and comprehensively. The data processing module can pre-process these data efficiently to form a standardized data set.

[0018] The intelligent analysis module uses deep neural networks to conduct in-depth analysis of standardized data sets, which can more accurately identify potential group buying opportunities and optimization solutions, thereby improving the efficiency and success rate of group buying.

[0019] The group buying optimization module can determine the preliminary group buying combination based on the results of intelligent analysis, and use dynamic adjustment factors to adjust the cost savings and transportation time prediction to ensure that the obtained group buying combination is optimal in terms of cost and timeliness, thereby generating more practical group buying suggestions.

[0020] The user interaction module can clearly display group buying suggestions to users and receive user feedback and instructions, allowing users to intuitively understand group buying plans and adjust and optimize them according to actual needs, greatly improving the user interaction experience.

[0021] The execution module can automatically complete operations such as space reservation, cargo transportation arrangement and fee settlement according to the group purchase plan confirmed by the user, which greatly improves the automation and work efficiency of group purchase operations, while reducing the possibility of human errors.

[0022] Through the intelligent group optimization system, enterprises can utilize transportation resources more effectively, reduce transportation costs, and improve transportation efficiency and cabin utilization. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 It is a schematic diagram of an intelligent grouping optimization system for bulk international freight based on data processing provided by an embodiment of the present invention.

[0024] Figure 2 It is a flow chart of a method for intelligent grouping optimization of bulk international freight based on data processing provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0025] The exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0026] like Figure 1As shown, the embodiment of the present invention proposes an intelligent grouping optimization system for bulk international freight based on data processing, including: Data collection module, used to collect real-time data related to bulk international freight, including cargo source information, transportation demand, ship schedule information, space information and cost information; A data processing module for preprocessing data related to bulk international freight to form a standardized data set; Intelligent analysis module, which uses deep neural networks to analyze standardized data sets to identify potential group buying opportunities and optimization solutions; The group purchase optimization module is used to determine preliminary group purchase combinations based on potential group purchase opportunities and optimization solutions; for each preliminary group purchase combination, the dynamic adjustment factor is used to adjust its cost savings and transportation time prediction to obtain an optimized group purchase combination; based on the optimized group purchase combination, the final group purchase suggestion is generated; The user interaction module is used to display group buying suggestions to users and receive user feedback and instructions to adjust and optimize group buying plans; The execution module is used to automatically execute the group buying operation according to the group buying plan confirmed by the user, including space reservation, cargo transportation arrangement and fee settlement.

[0027] In an embodiment of the present invention, through the data collection module, the system can collect various types of data related to bulk international freight in real time and comprehensively. The data processing module can efficiently pre-process these data to form a standardized data set. The intelligent analysis module uses a deep neural network to conduct an in-depth analysis of the standardized data set, and can more accurately identify potential group opportunities and optimization solutions, thereby improving the efficiency and success rate of group purchases. The group optimization module can determine the preliminary group combination based on the results of the intelligent analysis, and use dynamic adjustment factors to adjust the cost savings and transportation time prediction to ensure that the obtained group combination is optimal in both cost and time, thereby generating more practical group suggestions. The user interaction module can clearly display group suggestions to users and receive user feedback and instructions, so that users can intuitively understand the group plan and adjust and optimize it according to actual needs, greatly improving the user's interactive experience. The execution module can automatically complete operations such as space reservation, cargo transportation arrangement and fee settlement according to the group plan confirmed by the user, greatly improving the automation and work efficiency of group operations, while reducing the possibility of human errors. Through the intelligent group optimization system, enterprises can utilize transportation resources more effectively, reduce transportation costs, and improve transportation efficiency and cabin utilization.

[0028] When it comes to specific applications, the source of cargo information comes from major freight platforms, supply chain management systems or the company's internal ERP system; transportation needs are collected through direct communication with customers, online freight market platforms or the company's internal order management system; data is connected with shipping companies, ship agents or third-party logistics service platforms to obtain shipping schedule information; space information is provided by shipping companies or obtained through a dedicated space reservation system. Cost information includes transportation costs, insurance costs, tariffs, etc., and the data comes from financial systems, freight forwarders or related service providers. Use API (application programming interface) to exchange real-time data with various data sources, use web crawler technology to capture the required information from relevant websites or platforms, and set up RSS subscriptions or email notifications to receive the latest freight-related information. The collected data will first be stored in a temporary database for subsequent processing and analysis, and the data will usually be backed up to ensure data security and recoverability.

[0029] Data cleaning, removing duplicate data, ensuring data uniqueness, identifying and handling missing values, possibly through interpolation, deletion, or filling in according to business rules, detecting and handling outliers, such as extremely high or low values.

[0030] Convert data from different sources and formats into a standard format, encode categorical data, such as converting text descriptions into digital codes; standardize numerical data, such as using the z-score standardization method to make the data have zero mean and unit variance. Normalize the data and scale it to a specific range (such as between 0 and 1) to eliminate dimensional differences between different features. Extract features useful for analysis from the raw data, such as transportation distance, cargo weight, transportation time, etc.; after the above processing, the data will be integrated into a standardized data set. Through the collaborative work of the above two modules, the system can efficiently collect and process data related to bulk international freight.

[0031] In a preferred embodiment of the present invention, a standardized data set is analyzed using a deep neural network to identify potential group buying opportunities and optimization solutions, including: Determine the structure of the deep neural network model, including determining the configuration of the input layer, hidden layer, and output layer, specifically including: Determine the number of neurons in the input layer according to the number of features of the standardized data set. For example, if the data set contains 10 features, the input layer should be configured with 10 neurons. The number of hidden layers and the number of neurons in each layer need to be determined based on the complexity of the problem and the size of the data set. Usually, the best structure can be found by experimenting with different configurations. Hidden layers usually use activation functions such as ReLU to increase nonlinearity. The number of neurons in the output layer depends on the specific requirements of the prediction task. In group buying opportunity identification, the output is a probability value representing the possibility of a potential group buying opportunity, so a single neuron and a sigmoid activation function can be used.

[0032] The standardized dataset is divided into training set, validation set and test set, and the deep neural network model is trained using the training set. The usual ratio can be 70%:15%:15%. This is done to ensure that the model can be trained and evaluated on different data subsets, so as to more accurately evaluate its performance.

[0033] Adjust the weights of the deep neural network model through the gradient descent algorithm; Evaluate the performance of the deep neural network on the validation set, monitor overfitting during training, and adjust hyperparameters as needed; evaluate the performance of the trained deep neural network model on the test set to obtain evaluation results; optimize the deep neural network model based on the evaluation results, repeat the training and validation process to obtain the optimized deep neural network model, including: Read the validation set data from the partitioned dataset, which is new to the model during training; input the validation set data into the trained deep neural network model and calculate the prediction results through forward propagation; use evaluation indicators (such as accuracy, precision, recall, F1 score, etc.) to compare the prediction results with the true labels to evaluate the performance of the model on the validation set, and record the performance indicators after each validation for subsequent analysis and comparison. During the training process, calculate the loss function value on the validation set regularly (such as after each epoch). If the validation loss continues to rise for multiple consecutive epochs, while the training loss continues to fall, this may be a sign of overfitting. Set up a monitoring mechanism to stop training when the validation loss starts to rise to prevent the model from overfitting the training data. If overfitting is observed, regularization techniques (such as L1 regularization, L2 regularization, dropout, etc.) can be used to mitigate overfitting. According to the evaluation results on the validation set, analyze the reasons for the poor performance of the model, which may be due to improper settings of some hyperparameters. According to the analysis results, try to adjust the hyperparameters of the model (such as learning rate, batch size, number of hidden layer neurons, etc.) to seek better performance, record the situation of each hyperparameter adjustment and the corresponding validation set performance, so as to compare and select the best configuration later; read the test set data, which are not seen by the model in the entire training and optimization process, input the test set data into the trained and possibly optimized model, obtain the prediction results, and use the same evaluation indicators as the validation set to evaluate the prediction results on the test set to obtain the performance of the model on unknown data; analyze the evaluation results on the test set, find out the possible problems and deficiencies of the model, and adjust and optimize the model structure, training strategy or hyperparameters for the problems found, including increasing or decreasing the number of network layers, changing the activation function, adjusting the learning rate decay strategy, etc. Repeat the training and verification process with the optimized model configuration to verify the effectiveness of the optimization measures. The above optimization, training and verification steps can be repeated many times until satisfactory model performance is achieved to obtain the optimized deep neural network model.

[0034] The new standardized data is predicted by the optimized deep neural network model to identify potential group buying opportunities; based on the potential group buying opportunities, the cost saving and efficiency improvement potential between different source combinations are analyzed, and group buying opportunities with actual feasibility are screened, specifically including: inputting the new standardized data into the optimized deep neural network model, performing forward propagation calculations, and obtaining prediction results, which are expressed as the probability of potential group buying opportunities between different source combinations; parsing the prediction results, setting appropriate thresholds to determine which source combinations have higher group buying potential, and marking high-potential source combinations as potential group buying opportunities; for each potential group buying opportunity, collecting relevant cost data and efficiency indicators, such as transportation cost, time cost, warehousing cost, etc.; by comparing and analyzing the cost and efficiency changes of different source combinations before and after group buying, the potential for cost saving and efficiency improvement is calculated; based on actual conditions such as business rules, resource constraints, and market demand, potential group buying opportunities are screened to exclude those group buying opportunities that have cost saving and efficiency improvement potential in theory but are not feasible in actual operation.

[0035] Based on the group buying opportunities that are actually feasible, generate specific group buying optimization plans, including: for the group buying opportunities that are actually feasible after screening, formulate specific group buying optimization plans, and the plans should include details such as the selection of source combinations, transportation route planning, cabin booking strategies, and time arrangements.

[0036] In an embodiment of the present invention, a deep neural network is able to learn complex patterns and relationships in the data, thereby more accurately predicting potential group opportunities. This accuracy helps companies make smarter decisions and reduce unnecessary risks and costs. By analyzing the cost savings and efficiency improvement potential between different combinations of sources of goods, companies can more effectively allocate resources, such as ship space, transportation routes, and personnel. This helps to maximize the use of resources and improve overall operational efficiency. The group optimization solution can help companies reduce the transportation cost of a single source of goods and achieve economies of scale by merging the transportation needs of multiple sources of goods; by optimizing the group solution, companies can provide faster and more reliable freight services to meet customer expectations and needs. The deep neural network model can analyze new standardized data in real time, providing companies with instant group opportunities and optimization suggestions. By optimizing transportation efficiency and reducing costs, the group solution helps reduce unnecessary energy consumption and carbon emissions.

[0037] In a preferred embodiment of the present invention, the weights of the deep neural network model are adjusted by a gradient descent algorithm, including: The weight matrix and bias vector are randomly initialized for each layer of the deep neural network model, specifically including: for each layer of the deep neural network model (including input layer, hidden layer and output layer), the weight matrix W and the bias vector b are randomly initialized, the size of the weight matrix W is usually set to the number of neurons in the current layer multiplied by the number of neurons in the previous layer, and the size of the bias vector b is set to the number of neurons in the current layer; different methods can be used for initialization, such as zero initialization, random decimal value initialization (such as using normal distribution or uniform distribution) or specific initialization strategies (such as He initialization, Xavier initialization, etc.).

[0038] Select a training sample as input and forward propagate it through the deep neural network model. At each layer, calculate the dot product of the input and the weight, add the bias, and then get the corresponding output data through the activation function. Use the output data as the input of the next layer, and so on, until the prediction result of the final output layer is obtained. Specifically, select a training sample as input and pass it into the input layer of the deep neural network model. For each layer (starting from the input layer and propagating to the output layer layer by layer), perform the following operations: Calculate the weighted input of the current layer: z = W × a p +b, where a p is the output of the previous layer (or the input of the input layer), W and b are the weight and bias of the current layer; the weighted input z is passed through an activation function (such as ReLU, sigmoid, etc.) to obtain the output a of the current layer; the output a of the current layer is used as the input of the next layer, and so on, until the prediction result of the final output layer is obtained.

[0039] The mean square error is used to calculate the difference between the predicted value of the deep neural network model and the true value, specifically including: using the mean square error (MSE) to calculate the difference between the predicted value of the deep neural network model and the true value; for a single training sample, the mean square error The calculation formula is: ,in, is the predicted value of the model, is the true value, is the number of neurons in the output layer (or the dimension of the predicted value); in batch training, the errors of multiple training examples are averaged.

[0040] The gradient of the output layer with respect to the weight is calculated based on the mean square error. Using the chain rule, the gradient of each layer with respect to the weight is calculated back layer by layer from the output layer. The weight and bias of each layer are updated based on the calculated gradient and the set learning rate. Specifically, it includes: Calculate the gradient of the output layer with respect to the weights based on the mean square error or other loss function. For the output layer, the calculation of the gradient usually involves the derivative of the loss function with respect to the weighted input and the derivative of the activation function; use the chain rule to calculate the gradient of each layer with respect to the weights, starting from the output layer and working back layer by layer. This involves calculating the outer product of the gradient of the current layer with the output of the previous layer, and multiplying it by the derivative of the activation function of the previous layer; for each layer, the gradient of the weight matrix W and the bias vector b needs to be calculated, which will be used for subsequent weight updates. Update the weight matrix W and the bias vector b of each layer based on the calculated gradient and the set learning rate.

[0041] The update formula is usually: W2 = W-learning_rate × gradient_W and b2 = b-learning_rate × gradient_b, where gradient_W and gradient_b are the gradients with respect to weights and biases. The learning rate is a positive number that controls the step size of the weight update. A larger learning rate may lead to unstable training, while a smaller learning rate may lead to slow training. The updated weights and biases will be used in the forward propagation process of the next training iteration. By continuously repeating the above steps (i.e., the iterative training process), the weights and biases of the deep neural network model will gradually adjust to minimize the difference between the predicted value and the true value, thereby improving the prediction performance of the model.

[0042] In an embodiment of the present invention, the gradient descent algorithm continuously adjusts the network weights so that the network gradually "learns" and "grows", so that the task can be completed more accurately. This optimization process helps to reduce the gap between the model prediction and the actual value and improve the prediction accuracy. By calculating the mean square error to measure the difference between the model prediction value and the true value, and using the gradient descent algorithm to minimize this error, the training error of the model can be effectively reduced. As the iteration proceeds, the model will gradually converge to the vicinity of the optimal solution, thereby achieving better performance. By optimizing the weights, the deep neural network model can better capture the potential laws and features in the data, thereby improving its generalization ability. This means that the model can also make more accurate predictions when facing new data. By reasonably selecting the learning rate, the gradient descent algorithm can speed up the convergence speed as much as possible while ensuring the convergence stability. This helps to reduce training time and improve the training efficiency of the model. The gradient descent algorithm takes into account the noise and outliers in the data during the optimization process, thereby enhancing the robustness of the model to a certain extent. This enables the model to maintain good prediction performance when facing complex or uncertain data.

[0043] In a preferred embodiment of the present invention, based on potential group buying opportunities and optimization schemes, a preliminary group buying combination is determined, including: The number of individuals in the population is set, each individual represents a group combination, and the group combination is a list containing different source identifiers; the list is randomly generated so that the sources in each list meet the basic compatibility requirements until the preset conditions are met to form an initial population, specifically including: determining the number of individuals in the population, each individual (i.e., group combination) is represented by a list, the elements in the list are source identifiers, and these identifiers uniquely correspond to different sources; randomly generating lists, and continuously randomly generating lists until the preset population size is reached, and all generated lists that meet the conditions are combined to form an initial population.

[0044] Construct a fitness function, apply the fitness function to each individual in the initial population, and calculate the fitness value of each individual. The calculation formula of the fitness function is: ; in, Indicates The demand matching degree of a product indicates the degree to which the product meets the needs of the target user group; Indicates The estimated sales volume of the product in the group purchase combination; Indicates The current inventory quantity of the product; Indicates The unit cost of a commodity; represents an indicator variable. If If items are selected, ,otherwise ; , and represents the weight coefficient; is the standard deviation of demand matching; is the standard deviation of inventory utilization; is the standard deviation of the cost of goods; is the mean of the demand matching degree; is the mean of the inventory utilization ratio; It is The average cost of each product; It is the index of the product, indicating the number of each product in the group purchase combination; The value range is from 1 to , which means every product in the group purchase; is the total number of products, indicating how many products are included in the group purchase combination; ; in, It is The historical purchase volume of each product; It is Search volume for products; and is the weight coefficient; is the weight coefficient, It is The sales growth rate of a product.

[0045] According to the fitness value of the individual, the roulette wheel selection method is used to select some individuals from the population as parents, the selected parent individuals are paired, and single-point crossover is used to generate new offspring individuals; the generated offspring individuals are randomly mutated, and the parents, offspring and mutated individuals are combined into a new population, and the crossover and mutation operations are repeated until the termination condition is met, the iteration is stopped, and the corresponding individuals in the current population are output as the final solution, which specifically includes: according to the fitness value of the individual, the roulette wheel selection method (or other selection algorithms) is used to select some individuals from the current population as parents. The roulette wheel selection method is a selection method based on fitness ratio, and individuals with high fitness have a greater probability of being selected; the selected parent individuals are paired, and single-point crossover is used to generate new offspring individuals. Single-point crossover refers to randomly selecting a point in the list of parent individuals and then exchanging the gene fragments after the point; random mutation operations are performed on the generated offspring individuals to increase the diversity of the population. Mutation may involve randomly changing one or more source identifiers in the list; combining the parent, offspring, and mutated individuals to form a new population; repeating the crossover and mutation operations to generate a new population, and calculating the fitness value of the new population; checking after each iteration whether the termination condition is met (such as reaching the maximum number of iterations, the fitness value reaches a preset threshold, etc.), once the termination condition is met, stop the iteration process, and output the individual with the highest (or lowest, depending on the optimization goal) fitness value in the current population as the final solution.

[0046] Convert the final solution into the actual grouping combination, specifically including: decoding the final solution (a list containing source identifiers) into the actual grouping combination, that is, mapping the identifiers back to the actual source information.

[0047] In an embodiment of the present invention, by optimizing the group purchase combination through an intelligent algorithm, a combination method that meets the compatibility requirements and has the greatest benefit can be found more quickly, thereby improving the efficiency and success rate of group purchase. The optimized group purchase combination can match user needs more accurately and improve user satisfaction. Users can more conveniently participate in group purchase activities that meet their needs and enjoy more favorable prices and better shopping experiences. Reasonable group purchase combinations help reduce the operating costs of merchants. Through algorithm optimization, unnecessary logistics costs and inventory backlogs can be reduced, inventory turnover can be improved, and overall operating costs can be reduced. The optimized group purchase combination can stimulate consumers' desire to buy and increase sales. By finding the most attractive product combination through an intelligent algorithm, more users can be attracted to participate in group purchases, thereby improving sales performance. Through a reasonable group purchase combination, limited supply resources can be better utilized and waste can be reduced. The optimized combination scheme can ensure the efficient use of supply and avoid waste of resources. The method can adapt to different market demands and changes and quickly adjust the group purchase combination strategy. When market demand changes, the algorithm can quickly recalculate the optimal combination, allowing merchants to respond quickly to market changes. The group purchase combination optimization method based on an intelligent algorithm has good scalability and maintainability. As the business grows and the supply of goods increases, this method can be easily expanded to a larger scale and is easy to maintain and update.

[0048] In a preferred embodiment of the present invention, the dynamic adjustment factor The calculation formula is: ; in, Indicates The cost of each group purchase combination is calculated based on the actual procurement, production and transportation costs. For each group purchase combination, all relevant costs need to be listed in detail and summed up to get the total cost; Indicates The cost of a group purchase combination under the baseline situation, which refers to the normal cost of the group purchase combination in the absence of special circumstances (such as promotional activities, market fluctuations, etc.), which is determined through historical data; Indicates the number of group combinations; Indicates the shipping time of the group purchase combination, which is determined based on the data of the logistics provider or historical transportation data. It indicates the average time from shipment to receipt; Indicates the transportation time of the benchmark group purchase combination, which refers to the transportation time under normal circumstances without the influence of special factors. This can be obtained by analyzing historical transportation data; An adjustment factor that represents the effect of time on cost; Indicates The shipping time of each group purchase combination; Indicates The shipping time of each group purchase combination under the benchmark situation; Indicates the coefficient of impact of cost on transportation time; and is the weight coefficient.

[0049] In an embodiment of the present invention, the formula allows the weights of cost and transportation time, as well as the degree of mutual influence between them, to be dynamically adjusted according to actual conditions. This enables decision makers to respond to market changes more flexibly while ensuring the accuracy of decisions. By considering the cost and transportation time of group purchase combinations, the formula helps companies allocate resources more reasonably. For example, when the cost of a group purchase combination is high but the transportation time requirement is also high, the strategy can be adjusted according to the formula to ensure that the cost is reduced as much as possible while meeting the time requirements. By considering the impact of transportation time on cost, and the reverse impact of cost on transportation time, it helps companies to improve transportation time while ensuring cost-effectiveness, thereby improving customer satisfaction. Companies that can make decisions quickly and accurately and optimize resource allocation are often more competitive.

[0050] In a preferred embodiment of the present invention, for each preliminary group purchase combination, a dynamic adjustment factor is used to adjust its cost savings and transportation time efficiency prediction to obtain an optimized group purchase combination, including: For each preliminary group purchase combination, calculate its original cost savings forecast value, that is, the difference between the cost of the group purchase combination and the benchmark cost, including: Calculate the original transportation time forecast of the preliminary group combination, that is, the difference between the estimated transportation time of the group combination and the benchmark transportation time, including: extracting the cost data of each preliminary group combination and baseline cost data For each group combination, perform the calculation This gives a raw forecast of cost savings, which represents the amount of cost savings that the group buying combination is expected to achieve compared to the baseline scenario.

[0051] Multiply the original cost savings forecast by the dynamic adjustment factor to get the adjusted cost savings forecast, including: Retrieving the estimated shipping time for each preliminary group combination and benchmark transit time For each group combination, perform the calculation This is used to obtain the original predicted value of shipping time, which reflects the expected improvement in shipping time of the group purchase combination compared to the baseline situation.

[0052] The original transportation time forecast is multiplied by the dynamic adjustment factor to obtain the adjusted transportation time forecast value, which includes: determining the dynamic adjustment factor used to adjust the cost savings forecast value The original predicted cost savings for each group purchase combination is multiplied by the dynamic adjustment factor, and the adjusted cost savings prediction value is: ; The adjusted cost saving forecast value and the adjusted transportation time forecast value are combined to form a comprehensive forecast result, specifically including: multiplying the original transportation time forecast value of each group combination by the corresponding dynamic adjustment factor to obtain the adjusted transportation time forecast value; Based on the comprehensive prediction results, each preliminary group purchase combination is comprehensively evaluated to obtain an evaluation result; according to the evaluation result, the group purchase combination is optimized to obtain an optimized group purchase combination, which specifically includes: Create a new data set to store the adjusted cost savings forecast and the adjusted shipping time forecast, and standardize these data. Standardization can be done using different methods, such as Z-score standardization, Min-Max standardization, etc. The purpose is to convert indicators of different dimensions into the same comparable range. Determine a cost savings threshold, which represents the minimum cost savings level that the group purchase combination needs to achieve in order to be considered good performance; use this cost savings threshold as a fixed reference point for subsequent comparison with the comprehensive forecast results.

[0053] Compare the adjusted cost savings forecast of each group purchase combination with the set cost savings threshold; sort all group purchase combinations based on the comparison results, and the sorting can be based on the amount of cost savings; from the sorted list, select the best performing group purchase combinations, which not only meet the cost savings threshold requirements, but also show good performance in terms of transportation timeliness. Identify those group purchase combinations with poor performance, that is, those that do not meet the cost savings threshold or perform poorly in terms of transportation timeliness, and conduct in-depth cause analysis for these poor performance combinations. Possible reasons include unreasonable group purchase rules, uneconomical product combinations, and inefficient logistics paths. Based on the above cause analysis, formulate specific optimization strategies for each poorly performing group purchase combination. For example, adjust the number of group purchases or price discounts to attract more participants and reduce costs; recombining products to improve overall sales efficiency and transportation efficiency; optimize logistics paths to reduce transportation time and costs, etc. Apply these optimization strategies to actual group purchase combination management, and after optimization, generate new and improved group purchase combination solutions that can achieve a better balance between cost savings and transportation timeliness, thereby improving overall business performance.

[0054] In an embodiment of the present invention, the cost savings and transportation time efficiency of each group purchase combination can be predicted more accurately by dynamically adjusting the factors. This adjustment takes into account various dynamic factors in actual operations, thereby improving the accuracy of the prediction. The adjusted cost savings prediction value and transportation time efficiency prediction value can provide enterprises with more comprehensive information, and accurate transportation time efficiency prediction can help enterprises better manage customer expectations and reduce customer complaints caused by logistics delays. At the same time, by optimizing cost savings, enterprises can provide more competitive prices, thereby improving customer satisfaction and loyalty. The optimized group purchase combination can make more efficient use of enterprise resources, including inventory, manpower and logistics resources. This helps to reduce operating costs and improve overall operational efficiency. By comprehensively considering cost savings and transportation time efficiency, enterprises can better balance risks and benefits. The adjustment factor helps to identify and reduce potential risk points, such as cost overruns or transportation delays.

[0055] In a preferred embodiment of the present invention, a final group buying suggestion is generated based on the optimized group buying combination, including: For each optimized group buying combination, the weighted score is obtained by multiplying the adjusted cost savings forecast value and the adjusted transportation time forecast value by their respective weights and adding them together, which specifically includes: determining the data source containing historical data on cost savings and transportation time, which comes from past group buying activity records; extracting relevant historical data from the data source, including the cost savings value and transportation time value of each group buying activity; cleaning and preprocessing the data to ensure the accuracy and consistency of the data and exclude outliers or erroneous data.

[0056] Take cost savings and transportation time as independent variables, and the comprehensive effect of group buying activities (such as sales) as the dependent variable. Use the least squares method to construct a linear regression model to explore the impact of cost savings and transportation time on the comprehensive effect of group buying activities. Fit the model with historical data to estimate the coefficients of cost savings and transportation time (i.e., the preliminary estimate of the weight value). Use statistical methods (such as t-tests) to verify the model to ensure the significance and reliability of the model; check the residuals of the model (the difference between the actual value and the predicted value) to evaluate the degree of fit of the model, and interpret the coefficients of cost savings and transportation time as their weight values ​​for the comprehensive effect of group buying activities based on the final model. Among them, the calculation formula of the linear regression model is: Y=β0+β1×X1+β2×X2+ε; Among them, Y is the dependent variable, representing the comprehensive effect of group buying activities (such as sales); X1 is the first independent variable, representing cost savings; X2 is the second independent variable, representing transportation time; β0 is the intercept term, which represents the expected value of the dependent variable when all independent variables are 0; β1 and β2 are regression coefficients, respectively representing the influence of cost savings and transportation time on the comprehensive effect of group buying activities; ε is the error term.

[0057] The weighted scores are used as the sorting basis to sort the optimized group buying combinations from high to low to obtain a sorted group buying combination list, specifically including: using quick sort to sort the weighted scores from high to low, and after the sorting is completed, a group buying combination list arranged from high to low according to the weighted scores is obtained.

[0058] In the sorted group purchase combination list, filter according to the preset filtering conditions, including: evaluate the current market demand and trends, understand consumer preferences, price sensitivity and competitor dynamics; determine the key indicators of the filtering conditions based on the market analysis results, such as the cost saving threshold, the minimum requirement for transportation time, product type or brand restrictions, etc.; set filtering conditions, such as cost savings must reach a certain percentage, transportation time must not exceed a certain number of days, etc., and convert these conditions into operational judgment criteria for subsequent procedural processing. Check the group purchase combinations in the list one by one, and for each combination, extract its relevant data indicators, such as cost saving value, transportation time value, etc.; compare the data indicators of each combination with the filtering conditions. If all the indicators of a combination meet the filtering conditions, mark it as a group purchase combination that meets the requirements; otherwise, remove it from the list. After the above steps, a sub-list of group purchase combinations that meet all the preset conditions will be obtained.

[0059] From the filtered group purchase combinations, a corresponding combination is selected as the final group purchase suggestion according to the weighted score, specifically including: in the filtered group purchase combination sublist, searching for the combination with the highest weighted score, and selecting the group purchase combination with the highest weighted score as the final group purchase suggestion.

[0060] In the embodiment of the present invention, the weighted scores and ranking results provide enterprises with clear decision support, helping them to quickly identify group purchase combinations with excellent performance. Through the screening and selection process, enterprises can allocate resources more efficiently and focus on promoting and implementing group purchase activities with higher potential and returns. The optimized group purchase combination performs better in terms of cost savings and transportation timeliness, which helps to improve customer satisfaction.

[0061] In an embodiment of the present invention, after the final group purchase suggestions are generated, the user interaction module will display these suggestions to the user in an intuitive and easy-to-understand manner. This is usually implemented through a graphical user interface (GUI), such as a web page or a mobile application. The displayed content includes key information such as the product details, price, expected cost savings, and transportation time of the group purchase combination. The user interaction module provides an input interface, allowing users to provide feedback on the displayed group purchase suggestions. Users can provide feedback by clicking buttons, filling out forms, or using voice input. The module will capture the user's input in real time and convert it into instructions or data formats that the system can understand. Based on the user's feedback, the user interaction module will trigger the adjustment process of the group purchase plan. If the user is not satisfied with the suggestion, the system can provide modification options to allow the user to customize the group purchase conditions.

[0062] Once the user confirms the group purchase plan through the interactive module, the plan will be passed to the execution module, which will parse the various parameters and instructions in the plan to prepare for subsequent operations. According to the cargo transportation requirements in the group purchase plan, the execution module will automatically communicate with the space supplier to book the required space, which may involve negotiation and comparison with multiple suppliers to ensure the best space price and service quality. After the reservation is successful, the execution module will record the reservation details and update the system status information.

[0063] The execution module will formulate a detailed cargo transportation plan based on the shipping and receiving addresses, transportation time and other requirements in the group purchase plan. This includes selecting the appropriate mode of transportation (such as sea, air, land, etc.), arranging loading and unloading times, and monitoring the real-time location of the goods. The execution module will work closely with logistics partners to ensure that the goods are delivered to the destination on time and safely.

[0064] During the cargo transportation process, the execution module will track and record all relevant fees, such as space charges, transportation fees, insurance premiums, etc. Once the cargo transportation is completed, the execution module will automatically trigger the fee settlement process, and conduct reconciliation and payment operations with suppliers and logistics partners. After the settlement is completed, the execution module will generate a detailed expense report for users to review and audit.

[0065] like Figure 2 As shown, an embodiment of the present invention further provides a method for intelligent grouping optimization of bulk international freight based on data processing, comprising: Real-time collection of data related to bulk international freight, including cargo source information, transportation demand, ship schedule information, space information and cost information; Preprocessing of data related to bulk international shipping to form a standardized data set; Use deep neural networks to analyze standardized data sets to identify potential group buying opportunities and optimization solutions; Based on potential group buying opportunities and optimization plans, determine preliminary group buying combinations; for each preliminary group buying combination, use dynamic adjustment factors to adjust its cost savings and transportation time prediction to obtain an optimized group buying combination; based on the optimized group buying combination, generate the final group buying suggestion; Show group buying suggestions to users and receive feedback and instructions from users to adjust and optimize group buying plans; According to the group buying plan confirmed by the user, the group buying operation is automatically executed, including space reservation, cargo transportation arrangement and fee settlement.

[0066] It should be noted that the method is a method corresponding to the above-mentioned system, and all implementation methods in the above-mentioned method embodiment are applicable to this embodiment and can achieve the same technical effect.

[0067] The embodiment of the present invention further provides a computing device, including: a processor, a memory storing a computer program, and when the computer program is executed by the processor, the system described above is executed. All implementation methods in the above system embodiment are applicable to this embodiment and can achieve the same technical effect.

[0068] The embodiment of the present invention further provides a computer-readable storage medium storing instructions, which, when executed on a computer, enable the computer to execute the system described above. All implementations in the above system embodiment are applicable to this embodiment and can achieve the same technical effect.

[0069] The above is a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. An intelligent grouping optimization system for bulk international freight based on data processing, characterized in that: include: Data collection module, used to collect real-time data related to bulk international freight, including cargo source information, transportation demand, ship schedule information, space information and cost information; A data processing module for preprocessing data related to bulk international freight to form a standardized data set; Intelligent analysis module, which uses deep neural networks to analyze standardized data sets to identify potential group buying opportunities and optimization solutions; The group purchase optimization module is used to determine preliminary group purchase combinations based on potential group purchase opportunities and optimization solutions; for each preliminary group purchase combination, the dynamic adjustment factor is used to adjust its cost savings and transportation time prediction to obtain an optimized group purchase combination; based on the optimized group purchase combination, the final group purchase suggestion is generated; The user interaction module is used to display group buying suggestions to users and receive user feedback and instructions to adjust and optimize group buying plans; The execution module is used to automatically execute the group buying operation according to the group buying plan confirmed by the user, including space reservation, cargo transportation arrangement and fee settlement.

2. According to claim 1, a data processing-based intelligent group optimization system for bulk international freight transport, characterized in that: Use deep neural networks to analyze standardized data sets to identify potential group buying opportunities and optimization solutions, including: Determine the structure of the deep neural network model, including determining the configuration of the input layer, hidden layer, and output layer; Divide the standardized data set into training set, validation set and test set, and use the training set to train the deep neural network model; adjust the weight of the deep neural network model through the gradient descent algorithm; evaluate the performance of the deep neural network on the validation set, monitor the overfitting during the training process, and adjust the hyperparameters as needed; evaluate the performance of the trained deep neural network model on the test set to obtain the evaluation results; optimize the deep neural network model according to the evaluation results, and repeat the training and validation process to obtain the optimized deep neural network model; Use the optimized deep neural network model to predict new standardized data to identify potential group buying opportunities; analyze the cost savings and efficiency improvement potential between different source combinations based on potential group buying opportunities, and screen practical and feasible group buying opportunities; Generate specific group buying optimization plans based on actual feasible group buying opportunities.

3. According to claim 2, a data processing-based intelligent group optimization system for bulk international freight transport, characterized in that: Adjust the weights of the deep neural network model through the gradient descent algorithm, including: Randomly initialize the weight matrix and bias vector for each layer of the deep neural network model; Select a training sample as input and forward propagate it through the deep neural network model; At each layer, the dot product of the input and the weight is calculated, and the bias is added. Then the corresponding output data is obtained through the activation function, and the output data is used as the input of the next layer, and so on, until the prediction result of the final output layer is obtained; Use mean squared error to calculate the difference between the deep neural network model's prediction and the true value; The gradient of the output layer with respect to the weight is calculated according to the mean square error. Using the chain rule, the gradient of each layer with respect to the weight is calculated back layer by layer starting from the output layer. The weight and bias of each layer are updated according to the calculated gradient and the set learning rate.

4. The intelligent grouping optimization system for bulk cargo international freight based on data processing according to claim 3 is characterized in that: Based on potential group buying opportunities and optimization plans, determine the initial group buying combination, including: Set the number of individuals in the population, each individual represents a group combination, and the group combination is a list containing different source identifiers; randomly generate the list so that the sources in each list meet the basic compatibility requirements until the preset conditions are met to form the initial population; Construct a fitness function, apply the fitness function to each individual in the initial population, and calculate the fitness value of each individual; According to the fitness value of the individual, a roulette wheel selection method is used to select some individuals from the population as parents, the selected parent individuals are paired, and a single-point crossover is used to generate new offspring individuals; the generated offspring individuals are randomly mutated, and the parents, offspring and mutated individuals are combined into a new population, and the crossover and mutation operations are repeated until the termination condition is met, the iteration is stopped, and the corresponding individuals in the current population are output as the final solution; Convert the final solution into the actual puzzle combination.

5. According to claim 4, a data processing-based intelligent group optimization system for bulk international freight transport, characterized in that: The calculation formula of the dynamic adjustment factor is: ； in, Indicates The cost of a group combination; Indicates The cost of a group purchase combination in the baseline case; Indicates the number of group combinations; Indicates the transportation time of the group purchase combination; Indicates the transportation time of the benchmark group purchase combination; An adjustment factor that represents the effect of time on cost; Indicates The shipping time of each group purchase combination; Indicates The shipping time of each group purchase combination under the benchmark situation; Indicates the coefficient of impact of cost on transportation time; and is the weight coefficient.

6. The intelligent grouping optimization system for bulk cargo international freight based on data processing according to claim 5 is characterized in that: For each preliminary grouping combination, use dynamic adjustment factors to adjust its cost savings and transportation time prediction to obtain an optimized grouping combination, including: For each preliminary group purchase combination, calculate its original cost savings forecast value, that is, the difference between the cost of the group purchase combination and the benchmark cost; Calculate the original transportation time forecast of the preliminary group combination, that is, the difference between the estimated transportation time of the group combination and the benchmark transportation time; Multiply the original cost savings forecast by the dynamic adjustment factor to obtain the adjusted cost savings forecast; Multiply the original transportation time prediction by the dynamic adjustment factor to obtain the adjusted transportation time prediction value; Combine the adjusted cost savings forecast and the adjusted shipping time forecast to form a comprehensive forecast result; Based on the comprehensive prediction results, each preliminary group purchase combination is comprehensively evaluated to obtain the evaluation results; According to the evaluation results, the group buying combination is optimized to obtain the optimized group buying combination.

7. The intelligent grouping optimization system for bulk cargo international freight based on data processing according to claim 6 is characterized in that: Based on the optimized group buying combination, the final group buying suggestion is generated, including: For each optimized group purchase combination, the weighted score is obtained by multiplying the adjusted cost savings forecast value and the adjusted transportation time forecast value by their respective weights and adding them together; The weighted scores are used as the sorting basis to sort the optimized group purchase combinations from high to low to obtain a sorted group purchase combination list; In the sorted group combination list, filter according to the preset filtering conditions; From the screened group purchase combinations, the corresponding combination is selected as the final group purchase suggestion based on the weighted score.

8. A method for intelligent grouping optimization of bulk international freight based on data processing, characterized in that: The method is used to execute the system according to any one of claims 1 to 7, comprising: Real-time collection of data related to bulk international freight, including cargo source information, transportation demand, ship schedule information, space information and cost information; Preprocessing of data related to bulk international shipping to form a standardized data set; Use deep neural networks to analyze standardized data sets to identify potential group buying opportunities and optimization solutions; Based on potential group buying opportunities and optimization plans, determine preliminary group buying combinations; for each preliminary group buying combination, use dynamic adjustment factors to adjust its cost savings and transportation time prediction to obtain an optimized group buying combination; based on the optimized group buying combination, generate the final group buying suggestion; Show group buying suggestions to users and receive feedback and instructions from users to adjust and optimize group buying plans; According to the group buying plan confirmed by the user, the group buying operation is automatically executed, including space reservation, cargo transportation arrangement and fee settlement.

9. A computing device, characterized in that include: one or more processors; A storage device for storing one or more programs, when the one or more programs are executed by the one or more processors, the one or more processors implement the system according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a program, and when the program is executed by a processor, the system according to any one of claims 1 to 7 is implemented.

Images