Intelligent express order processing optimization system and method based on big data analysis

Abstract

The application belongs to the technical field of logistics management, and discloses an intelligent express order processing optimization system and method based on big data analysis, which comprises the following steps: collecting data related to express orders in a target express company to obtain corresponding logistics order information; the logistics order information comprises express order information and order auxiliary data; performing data processing on the collected logistics order information to obtain a corresponding order decision tree diagram; obtaining a to-be-processed order, combining the order decision tree diagram to formulate a corresponding initial order processing strategy for the corresponding to-be-processed order; performing strategy verification on the obtained initial order processing strategy, and taking corresponding feedback measures based on the strategy verification result; and efficient management and optimization of express order information are realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of logistics management, and more particularly, relates to an intelligent express order processing optimization system and method based on big data analysis. BACKGROUND

[0002] With the rapid development of e-commerce and the increasing globalization of trade, the express industry is facing unprecedented challenges and opportunities. Traditional order processing methods have been unable to meet the needs of modern logistics;

[0003] Compared with the prior art, the traditional order processing method has the problems of low efficiency, high error rate, difficulty in adapting to large-scale business demand, etc., and cannot fully consider the personalized needs of users; at the same time, it is difficult to meet the requirements of real-time and accuracy; and there are many problems such as lack of effective data processing means when processing complex logistics order information, which leads to inability to deeply mine the hidden information and rules behind the data, etc. In view of this, the present application provides an intelligent express order processing optimization system and method based on big data analysis to solve the above problems. SUMMARY

[0004] In order to overcome the above-mentioned defects of the prior art, in order to achieve the above-mentioned purpose, the present application provides the following technical scheme:

[0005] The intelligent express order processing optimization system based on big data analysis comprises:

[0006] The data acquisition module is used for acquiring data related to express orders in the target express company, and obtaining corresponding logistics order information; the logistics order information comprises express order information and order auxiliary data;

[0007] The data processing module is used for processing the collected logistics order information, and obtaining a corresponding order decision tree diagram;

[0008] The order processing module is used for obtaining a to-be-processed order, and formulating an initial order processing strategy for the corresponding to-be-processed order in combination with the order decision tree diagram;

[0009] The strategy verification module is used for verifying the initial order processing strategy obtained, and taking corresponding feedback measures based on the strategy verification result.

[0010] Further, the process of acquiring data related to express orders in the target express company and obtaining corresponding logistics order information comprises:

[0011] The data acquisition module is internally provided with a plurality of data acquisition nodes, which acquire data related to express orders stored in the management center of the target express company based on the data acquisition nodes, and obtain a plurality of express order information, including user data, order data and delivery data.

[0012] Meanwhile, the data acquisition nodes are also used for data acquisition from external data sources to obtain corresponding order auxiliary data; the external data sources include meteorological bureaus and social media.

[0013] The corresponding express order information and order auxiliary data corresponding to the order processing process are summarized to obtain corresponding logistics order information.

[0014] Further, the process of data processing of the collected logistics order information includes:

[0015] The collected logistics order data is pre-processed to obtain corresponding initial logistics information; the data preprocessing includes data cleaning and data integration;

[0016] The order data in the initial logistics information is obtained, and the corresponding order data is data filtered to obtain corresponding order index data;

[0017] The corresponding order index data is classified based on the pre-constructed index classification function to obtain corresponding order feature data; the order feature data includes first order feature and second order feature;

[0018] The time consumed by the corresponding order data based on the initial logistics information to obtain the processing process is obtained, which is marked as order processing time, and the mapping relationship between order processing time and order feature data is constructed;

[0019] The flowchart of the order processing is obtained, and each feature index in the order feature data is mapped to the corresponding order processing flowchart, and the correlation between different feature indexes is obtained based on it; wherein the feature index is the first order feature or the second order feature;

[0020] The number of times of each feature index appearing in the order data in different initial logistics information is obtained, and the frequency characteristic value corresponding to the corresponding feature index is obtained based on it;

[0021] The frequency characteristic values of the first order feature and the second order feature are sorted respectively, and from the root node, the corresponding first order feature and second order feature are added to the tree according to the sorting of the frequency characteristic value, to obtain corresponding first order feature and second order feature. The first index branch and the second index branch are obtained;

[0022] Further, the corresponding first order data and the second order data are summarized to obtain corresponding first order data.

[0023] Further, the corresponding order index data is classified based on the pre-constructed index classification function to obtain the corresponding order feature data, and the process includes:

[0024] The index classification function is defined as: ; in the formula, represents a normalization process; represents the regularization coefficient corresponding to the i-th order index data in the j-th initial order data, ; represents the regression coefficient of the i-th order index data; ; represents the regularization factor; represents the reference value of the i-th order index data; ; and is an integer, represents the total number of order data; represents the index of order data; > 0 and is an integer, represents the total number of order index data in the corresponding order data; represents the index of order index data; The regularization coefficient corresponding to each order index data in the corresponding order data is obtained, and compared with the pre-set regularization threshold range; If the regularization coefficient is greater than the maximum value of the regularization threshold range, the corresponding order index data is marked as the first order feature; if the regularization coefficient is within the regularization threshold range, the corresponding order index data is marked as the second order feature; if the regularization coefficient is less than the minimum value of the regularization threshold range, no other operation is performed; The corresponding first order feature and the second order feature are summarized to obtain the corresponding first order data;

[0025] Further, the first order data corresponding to the order data in all initial logistics information is obtained, and data merging is performed to obtain the corresponding order feature data.

[0026] Further, the corresponding first order data and the second order feature are summarized to obtain the corresponding first order data.

[0027] Further, the first order data corresponding to the order data in all initial logistics information is obtained, and data merging is performed to obtain the corresponding order feature data.

[0028] Further, the corresponding first order data and the second order feature are summarized to obtain the corresponding first order data.

[0029] ​​​Further, the obtaining process of the order decision tree diagram comprises:

[0030] Based on the obtaining process of the order bifurcation tree, the corresponding user bifurcation tree and the delivery bifurcation tree are obtained respectively; the mapping relationship between the order data, the delivery data, the user data and the corresponding order processing time is obtained respectively; the corresponding user bifurcation tree, the delivery bifurcation tree and the order bifurcation tree are merged to obtain the corresponding order decision tree diagram.

[0031] Further, the process of obtaining the to-be-processed order and formulating the corresponding initial order processing strategy for the corresponding to-be-processed order in combination with the order decision tree diagram comprises:

[0032] Based on the data acquisition node, the to-be-processed order information corresponding to the to-be-processed order constructed in the management platform is obtained; the to-be-processed order information comprises order data and user data;

[0033] Based on the to-be-processed order information, the corresponding order decision tree diagram is traversed, and node similar matching is performed to obtain the root node similar to the corresponding to-be-processed order data and user data in the corresponding to-be-processed order information;

[0034] Based on the mapping relationship between the corresponding root node and the order processing time, the order processing time of the corresponding to-be-processed order is predicted to obtain the corresponding expected order processing time;

[0035] The obtained to-be-processed order information and expected order processing time are input into the pre-constructed order decision model to obtain the corresponding initial order processing strategy.

[0036] Further, the construction process of the order decision model comprises:

[0037] The constructed order decision tree diagram is obtained, and based on the same, a plurality of training data sets are constructed, wherein the training data set is composed of a plurality of training subsets;

[0038] The network framework of the order decision model is defined as a convolutional neural network for learning the mapping relationship among the order data, the delivery data and the user data;

[0039] The basic framework of the convolutional neural network is an input layer, a hidden layer and an output layer; the input layer is used to receive the input training subset and perform linear transformation thereon to obtain a corresponding feature matrix;

[0040] The hidden layer is used to perform nonlinear transformation on the corresponding feature matrix, extract the features of the feature matrix, learn the rules and patterns in the data, and pass these features to the output layer for decision-making;

[0041] The output layer is configured to receive the output of the hidden layer, and make a decision on the input data according to the features extracted by the hidden layer and the rules and patterns in the learning data;

[0042] Define the objective function of the order decision model ; wherein, represents the order processing strategy at time t; expresses the expected value ; represents the discount factor at time t; and respectively represent the strategy state and the strategy action at time t; represents the reward function; ; represents the total length of the expected order processing time input;

[0043] Define the loss function of the order decision model as ; wherein, represents the label of the input th training subset, and is an integer, represents the total number of training subsets; represents the expected result of the output result corresponding to the th sample; based on the training data set, the corresponding order decision model is iteratively trained, and based on the optimizer continuously optimizes the model parameters of the corresponding order decision model in the iterative training process until the difference between the loss function in the order decision model and the corresponding objective function no longer decreases or changes for consecutive training subsets, the model parameters are saved, i.e. the construction process of the order decision model is completed; wherein, is a constant.

[0044] Further, the process of verifying the obtained initial order processing strategy includes:

[0045] Obtain the whole process of the corresponding order processing, and construct a corresponding order simulation model based thereon;

[0046] Input the obtained initial order processing strategy and the real-time collected order auxiliary data into the corresponding order simulation model, record the running process of the corresponding order simulation model, and obtain the corresponding simulation order information;

[0047] Obtain the whole process of the corresponding order processing, and divide it into detailed processes; obtain a plurality of order processing sub-processes;

[0048] Obtaining process data corresponding to different order sub-processes based on the simulation order information, and marking the process data as first-level parameters; marking sub-data in the corresponding process data as second-level parameters;

[0049] Constructing a fuzzy process evaluation interval, and setting a plurality of sub-intervals in the fuzzy process evaluation interval;

[0050] Obtaining the membership degrees of the corresponding second-level parameter data and the corresponding fuzzy process evaluation interval;

[0051] Obtaining the membership degrees between the corresponding second-level parameter data and all sub-intervals, and selecting the two sub-intervals with the highest membership degrees and , and marking the membership degrees corresponding to the corresponding sub-intervals as and ;

[0052] Further, obtaining the correlation parameters between the corresponding second-level parameters and the corresponding sub-intervals ;

[0053] Obtaining the correlation parameters between all second-level indicators in the corresponding first-level indicator and the corresponding fuzzy process evaluation interval, and constructing a corresponding second-level correlation matrix based thereon;

[0054] Obtaining the product of the weight row vector corresponding to each second-level parameter and the corresponding correlation matrix, obtaining the correlation factors between the corresponding first-level parameters and the corresponding fuzzy grades, and constructing a first-level correlation matrix based thereon;

[0055] Further, obtaining the product between the weight row vector corresponding to each first-level parameter and the corresponding first-level correlation matrix, and obtaining the strategy characteristic value corresponding to the corresponding initial order strategy;

[0056] Setting a characteristic threshold, and comparing the obtained strategy characteristic value with the corresponding characteristic threshold;

[0057] If the strategy characteristic value is not less than the characteristic threshold, the corresponding initial order processing strategy is taken as the optimal order processing strategy, a corresponding process execution instruction is generated based thereon, and instruction feedback is performed;

[0058] If the strategy characteristic value is less than the characteristic threshold, the sub-interval corresponding to each second-level parameter is obtained, and based on whether the upper limit value of the corresponding sub-interval is lower than a preset parameter threshold, if not, no other operation is performed; if yes, the corresponding second-level parameter is marked as a to-be-optimized parameter, and based on the order decision tree diagram, it is determined whether the corresponding to-be-optimized parameter is a characteristic indicator, if yes, the corresponding to-be-optimized indicator data is input into the corresponding order decision model as a constraint condition, a new initial order processing strategy is obtained, and the process is repeated; if not, the corresponding to-be-optimized indicator is fed back to the management platform for manual intervention.

[0059] Further, the membership degree of the secondary parameter data to the corresponding fuzzy process evaluation interval ; wherein, represents the first secondary parameter in the first primary parameter and the first sub-interval in the fuzzy process evaluation interval; and respectively represent the upper limit value and the lower limit value of the first sub-interval in the fuzzy process evaluation interval; represents the first secondary parameter in the first primary parameter; wherein, and is an integer, represents the total number of order sub-processes; and is an integer, represents the total number of secondary parameters in the corresponding order sub-process;

[0060] the association parameter between the corresponding secondary parameter and the corresponding sub-interval ; wherein, represents the association parameter between the first secondary parameter in the first primary parameter and the first sub-interval in the fuzzy process evaluation interval.

[0061] Further, the intelligent express order processing optimization method based on big data analysis comprises:

[0062] Step 1: Collecting data related to express orders in the target express company to obtain corresponding logistics order information; the logistics order information includes express order information and order auxiliary data;

[0063] Step 2: Data processing of the collected logistics order information to obtain a corresponding order decision tree diagram;

[0064] Step 3: Obtaining a to-be-processed order and combining the order decision tree diagram to formulate an initial order processing strategy for the corresponding to-be-processed order;

[0065] Step 4: Strategy verification of the obtained initial order processing strategy, and taking corresponding feedback measures based on the strategy verification result.

[0066] The technical effects and advantages of the intelligent express order processing optimization system and method based on big data analysis are:

[0067] ​1. Through complex data preprocessing and feature extraction on collected logistics order information, an order decision tree diagram is generated; and based on the construction of an index classification function, order index data is classified, effectively identifying the key factors affecting order processing efficiency;

[0068] 2. The construction of the order decision tree diagram can provide personalized initial order processing strategies for different order types and user needs, improve the efficiency of resource allocation, and reduce operating costs; at the same time, the effectiveness of the strategy is ensured by verifying the order processing strategy, and the strategy is adjusted according to the feedback to continuously improve the service quality. BRIEF DESCRIPTION OF DRAWINGS

[0069] Figure 1 The figure is a schematic diagram of the intelligent express order processing optimization system based on big data analysis of the present application;

[0070] Figure 2 The figure is a schematic diagram of the intelligent express order processing optimization method based on big data analysis of the present application. DETAILED DESCRIPTION

[0071] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative labor fall within the scope of the present application.

[0072] Embodiment 1

[0073] Please refer to Figure 1 The intelligent express order processing optimization system based on big data analysis of the present embodiment includes:

[0074] The data acquisition module is used for acquiring data related to express orders in the target express company, and obtaining corresponding logistics order information; the logistics order information includes express order information and order auxiliary data;

[0075] The data processing module is used for processing the collected logistics order information, and obtaining corresponding order decision tree diagrams;

[0076] The order processing module is used for obtaining orders to be processed, and formulating corresponding initial order processing strategies for corresponding orders to be processed in combination with the order decision tree diagram;

[0077] The strategy verification module is used for verifying the obtained initial order processing strategy, and taking corresponding feedback measures based on the strategy verification result;

[0078] The various modules are connected through wired and / or wireless means to achieve data transmission between the modules.

[0079] It should be further explained that, in the specific implementation process, the process of collecting data related to express orders in the target express company and obtaining corresponding logistics order information includes:

[0080] The data acquisition module is provided with a plurality of data acquisition nodes, the data acquisition nodes are connected with the management platform of the target express company through API interface or database, and the data related to express orders stored in the management platform of the target express company are collected based on the data acquisition nodes to obtain a plurality of express order information, including user data, order data and distribution data; wherein, the user data includes user preferences, user basic information, etc.; the order data includes commodity type, commodity quantity, etc.; the distribution data includes distribution route, distribution time, etc.

[0081] Meanwhile, the data acquisition nodes are also used for data acquisition from external data sources to obtain corresponding order auxiliary data; wherein, the external data sources include meteorological bureau and social media; the order auxiliary data includes weather forecast, traffic condition and other data related to express order processing efficiency;

[0082] The corresponding express order information and order auxiliary data corresponding to the order processing process are summarized to obtain corresponding logistics order information.

[0083] It should be further explained that, in the specific implementation process, the process of collecting the collected logistics order information for data processing to obtain corresponding order decision tree diagram includes:

[0084] The collected logistics order data is preprocessed to obtain corresponding initial logistics information; the data preprocessing includes data cleaning and data integration; wherein, the data cleaning refers to deleting the repeated data in the collected logistics order information, filling and correcting the missing part and abnormal data in the logistics order information; the data integration refers to data standardization processing of the logistics order information after data cleaning, mapping the corresponding logistics order data to the pre-set data format, facilitating subsequent data processing;

[0085] The order data in the initial logistics information is obtained, and the corresponding order data is screened to obtain corresponding order index data, which refers to the index data in the order data that can reflect the corresponding order processing efficiency from different aspects;

[0086] Further, the corresponding order index data is classified based on the pre-constructed index classification function to obtain corresponding order feature data.

[0087] The process of classifying the corresponding order indicator data based on a pre-built indicator classification function to obtain the corresponding order feature data includes:

[0088] Define the index classification function as follows: In the formula, This indicates normalization processing; Indicates the first Within the initial order data, the first The regularization coefficient corresponding to each order metric data, where... ; Indicates the first Regression coefficients for individual order metrics data; This represents the regularization factor, used for feature selection and dimensionality reduction; a larger value indicates a higher regularization factor. The value increases the penalty on the coefficient, causing more regularization coefficients to become zero; Indicates the first Within the initial order data, the first Individual order metrics data; Indicates the first Reference values ​​for individual order metrics; and It is an integer. This indicates the total number of order data; An index representing order data; >0 and It is an integer. This indicates the total number of order indicator data within the corresponding order data; This represents an index for order indicator data;

[0089] Taking a specific order data as an example, obtain the regularization coefficients corresponding to each order indicator data within the corresponding order data, and compare them with the pre-set regularization threshold range;

[0090] If the regularization coefficient is greater than the maximum value of the regularization threshold range, the corresponding order indicator data is marked as the first order feature; if the regularization coefficient is within the regularization threshold range, the corresponding order indicator data is marked as the second order feature; if the regularization coefficient is less than the minimum value of the regularization threshold range, no other operation is performed.

[0091] The first order features and second order features are summarized to obtain the corresponding first order data. The influence of the feature indicators on the order processing of the corresponding order sub-process is as follows: the first order features are greater than the second order features, and the larger the regularization coefficient of the corresponding feature indicator, the higher its influence.

[0092] Further, the first order data corresponding to all the initial logistics information order data is obtained, and data merging is performed to obtain corresponding order feature data; the order feature data includes first order features and second order features; wherein, during the data merging process, repeated first order features and second order features are removed;

[0093] Further, the time consumed by the corresponding order data in the processing process is obtained based on the initial logistics information, and is marked as order processing time, and a mapping relationship between the order processing time and the order feature data is constructed;

[0094] The flowchart of order processing is obtained, and each feature index in the order feature data is mapped to the corresponding order processing flowchart, and the correlation between different feature indexes is obtained based thereon; wherein, the feature index is the first order feature or the second order feature;

[0095] The number of times each feature index appears in the order data in different initial logistics information is obtained, and the frequency characteristic value corresponding to the corresponding feature index is obtained based thereon;

[0096] The frequency characteristic values of the first order features and the second order features are sorted respectively, and from the root node, the corresponding first order features and second order features are added to the tree according to the sorting of the frequency characteristic values, to obtain corresponding first index branches and second index branches;

[0097] Further, the corresponding first index branches and second index branches are merged based on the obtained correlation between the feature indexes to obtain a corresponding order bifurcation tree;

[0098] Further, based on the obtaining process of the order bifurcation tree, the corresponding user bifurcation tree and the distribution bifurcation tree are obtained respectively; wherein, the user data corresponds to the user bifurcation tree, and the user bifurcation tree can be used to reflect the preference and satisfaction of the corresponding user to the order processing process; it is helpful for express companies to develop more personalized marketing strategies and service solutions; the distribution data corresponds to the distribution bifurcation tree;

[0099] Further, the mapping relationship between the order data, the distribution data and the user data and the corresponding order processing time is obtained, the corresponding user bifurcation tree, the distribution bifurcation tree and the order bifurcation tree are merged, and a corresponding order decision tree diagram is obtained.

[0100] It needs to be further explained that, in the specific implementation process, the process of obtaining the to-be-processed order and formulating the corresponding initial order processing strategy for the corresponding to-be-processed order in combination with the order decision tree diagram includes:

[0101] The to-be-processed order information corresponding to the to-be-processed order constructed by the data acquisition node is obtained from the management platform; the to-be-processed order information includes order data and user data;

[0102] Based on the order information to be processed, the corresponding order decision tree diagram is traversed, and the root node similar to the corresponding order data to be processed and user data is obtained through node similarity matching; then, based on the mapping relationship between the corresponding root node and the order processing time, the order processing time of the corresponding order to be processed is predicted, and the expected order processing time is obtained;

[0103] Then, the obtained order information to be processed and the expected order processing time are input into the pre-constructed order decision model to obtain the corresponding initial order processing strategy; the order processing strategy refers to the resource allocation method and distribution plan involved in the allocation of manpower, vehicles, warehouse space and inventory during the corresponding order processing process;

[0104] It should be further pointed out that in the specific implementation process, the construction process of the order decision model includes:

[0105] The constructed order decision tree diagram is obtained, and based on it, a plurality of training data sets are constructed, wherein the training data set is composed of a plurality of training subsets, and the training subset can be used to reflect the mapping relationship between the corresponding order data, user data and distribution data corresponding to the feature data and the mapping relationship with the distribution time; wherein the feature data includes the first user feature, the second user feature and the first distribution feature, the second distribution feature corresponding to the corresponding user data and the distribution data; also includes the first order feature and the second order feature, wherein the acquisition process of the corresponding first user feature, the second user feature and the first distribution feature, the second distribution feature is the same as that of the corresponding first order feature and the second order feature, and the present application does not make too much elaboration;

[0106] The network framework of the order decision model is defined as a convolutional neural network, which is used to learn the mapping relationship between the order data, the distribution data and the user data;

[0107] The basic framework of the convolutional neural network is the input layer, the hidden layer and the output layer; the input layer is used to receive the input training subset and perform linear transformation thereon to obtain the corresponding feature matrix;

[0108] The hidden layer is used to perform nonlinear transformation on the corresponding feature matrix, extract the features of the feature matrix, learn the rules and patterns in the data, and pass these features to the output layer for decision-making;

[0109] The output layer is used to receive the output of the hidden layer, and according to the features extracted by the hidden layer and the rules and patterns in the learning data, the input data is decided;

[0110] The objective function of the order decision model is defined ; in the formula, represents The order processing strategy at the moment; The expression selects the expected value The maximum value; The expression The discount factor at the moment; And Respectively represent the policy state and the policy action at t moment; The expression represents the reward function; ; The expression represents the total length of the input expected order processing time;

[0111] The loss function of the order decision model is defined as ; In the formula, The input label of the th training subset, And is an integer, The total number of training subsets; The expression represents the expected result of the output result corresponding to the th sample; Based on the training data set, the corresponding order decision model is iteratively trained, and based on The optimizer continuously optimizes the model parameters of the corresponding order decision model in the iterative training process until the difference between the loss function in the order decision model and the corresponding target function is continuously The training subset no longer decreases or changes, and the model parameters are saved, that is, the construction process of the order decision model is completed; wherein, is a constant.

[0112] It needs to be further explained that in the specific implementation process, the process of verifying the obtained initial order processing strategy based on the strategy verification result and taking corresponding feedback measures includes:

[0113] Obtain the whole process of the corresponding order processing, and construct the corresponding order simulation model based on it; wherein, the order simulation model refers to the dynamic model of the logistics distribution center workflow order processing system created based on the Module library, which is used to explore the relationship between the order processing flow and the corresponding various strategies, so as to quickly find out the best optimization scheme in the corresponding order processing optimization; wherein, The construction process of the module library and the order simulation model is the prior art, and the present application will not be described in detail;

[0114] The obtained initial order processing strategy and the real-time collected order auxiliary data are input into the corresponding order simulation model, and the running process of the corresponding order simulation model is recorded. The simulation order information is obtained;

[0115] Obtain the whole process of the corresponding order processing, and finely divide the process; obtain several order processing sub-processes;

[0116] Further, based on the simulation order information, the process data corresponding to different order sub-processes is obtained respectively, and is marked as a first-level parameter; the sub-data in the corresponding process data is marked as a second-level parameter;

[0117] A fuzzy process evaluation interval is constructed, and several sub-intervals are arranged in the fuzzy process evaluation interval, which can be used to express the evaluation level of the corresponding process under different second-level parameters;

[0118] Further, the membership degrees of the corresponding second-level parameter data and the corresponding fuzzy process evaluation interval are obtained ; wherein, represents the membership degree between the first second-level parameter in the first first-level parameter and the first sub-interval in the fuzzy process evaluation interval; and respectively represent the upper limit value and the lower limit value of the first sub-interval in the fuzzy process evaluation interval; represents the first second-level parameter in the first first-level parameter; wherein, and are integers, represents the total number of order sub-processes; and are integers, represents the total number of second-level parameters in the corresponding order sub-process; The membership degrees between the corresponding second-level parameter data and all sub-intervals are obtained, and the two sub-intervals with the highest membership degrees are selected and

[0119] respectively, and the membership degrees corresponding to the corresponding sub-intervals are marked as and ; Further, the correlation parameters between the corresponding second-level parameter and the corresponding sub-interval are obtained; wherein,

[0120] represents the correlation parameter between the first second-level parameter in the first first-level parameter and the first sub-interval in the fuzzy process evaluation interval; wherein, The higher the correlation parameter is, the closer the data attribute of the corresponding second-level parameter is to the sub-interval ;

[0121] ​​​​​​​​​​Further, the correlation parameters between all the second-level indexes in the corresponding first-level index and the corresponding fuzzy process evaluation interval are obtained, and a corresponding second-level correlation matrix is constructed based on the correlation parameters;

[0122] The product of the weight row vector corresponding to each second-level parameter and the corresponding correlation matrix is obtained, the correlation factors between the corresponding first-level parameter and the corresponding fuzzy level are obtained, and a first-level correlation matrix is constructed based on the correlation factors;

[0123] Further, the product of the weight row vector corresponding to each first-level parameter and the corresponding first-level correlation matrix is obtained, the strategy characteristic value corresponding to the corresponding initial order strategy is obtained, the strategy characteristic value is used to measure the level variable characteristic value between the initial order processing strategy and the corresponding order processing evaluation, and indicates the degree to which the evaluation of the initial order processing strategy is biased towards a certain fuzzy process evaluation interval;

[0124] A characteristic threshold is set, and the obtained strategy characteristic value is compared with the corresponding characteristic threshold;

[0125] If the strategy characteristic value is not less than the characteristic threshold, the corresponding initial order processing strategy is taken as the optimal order processing strategy, the corresponding process execution instruction is generated based on the initial order processing strategy, and the process execution instruction is fed back to the workers or freight robots of the corresponding order sub-process; the process execution instruction includes the order processing strategy and the expected processing time interval of the corresponding order process;

[0126] If the strategy characteristic value is less than the characteristic threshold, the sub-interval corresponding to each second-level parameter is obtained, and based on whether the upper limit value of the corresponding sub-interval is lower than the preset parameter threshold, if not, no other operation is performed; if lower, the corresponding second-level parameter is marked as a to-be-optimized parameter, and whether the to-be-optimized parameter is a characteristic index is determined based on the order decision tree diagram, if it is a characteristic index, the corresponding to-be-optimized index data is input into the corresponding order decision model as a constraint condition (i.e. the initial order processing strategy generation process of the corresponding order decision model is intervened based on the to-be-optimized index data); a new initial order processing strategy is obtained, and the process is repeated; if it is not a characteristic index, the corresponding to-be-optimized index is fed back to the management platform, and the corresponding workers manually intervene in the corresponding original initial order processing strategy.

[0127] It should be further explained that in the specific implementation process, the weight row vector corresponding to the first-level parameter and the second-level parameter is set by the workers in the field according to the expert entropy weight algorithm; the specific setting process will not be described in detail in the present application;

[0128] The application realizes intelligent optimization of express orders through data acquisition, data processing, order processing and strategy verification modules; and through construction of an order decision tree diagram and an order decision model, the order processing time can be predicted and an initial order processing strategy can be formulated, the strategy verification module ensures the effectiveness and feasibility of the initial order processing strategy, and provides strong support for order processing of express companies.

[0129] Embodiment 2

[0130] Please refer to Figure 2 The embodiment does not describe some parts in detail, and the intelligent express order processing optimization method based on big data analysis is provided, including:

[0131] Step one: collect data related to express orders in a target express company to obtain corresponding logistics order information; the logistics order information includes express order information and order auxiliary data;

[0132] Step two: data processing is performed on the collected logistics order information to obtain a corresponding order decision tree diagram;

[0133] Step three: obtain an order to be processed, and formulate an initial order processing strategy for the corresponding order to be processed in combination with the order decision tree diagram;

[0134] Step four: strategy verification is performed on the obtained initial order processing strategy, and corresponding feedback measures are taken based on the strategy verification result.

[0135] Embodiment 3

[0136] The embodiment discloses an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor implements the running mode of the intelligent express order processing optimization system and method based on big data analysis provided above when executing the computer program.

[0137] Since the electronic device introduced in the embodiment is the electronic device used to implement the intelligent express order processing optimization system and method based on big data analysis in the embodiment, the specific implementation of the electronic device and its various changes can be understood by those skilled in the art based on the intelligent express order processing optimization system and method based on big data analysis introduced in the embodiment, so the implementation of the electronic device in the method in the embodiment will not be described in detail. As long as the electronic device used to implement the intelligent express order processing optimization system and method based on big data analysis in the embodiment is implemented by those skilled in the art, it belongs to the scope of the application.

[0138] The above formulas are all dimensionless values calculated, the formulas are obtained by collecting a large amount of data to simulate a formula of the most recent real situation, and preset parameters and threshold values in the formulas are set by a person skilled in the art according to actual conditions.

[0139] The above are only preferred embodiments of the present application, and the protection scope of the present application is not limited to the above-mentioned embodiments. Any technical solution falling within the concept of the present application shall fall within the protection scope of the present application. It should be noted that, for ordinary technical users in the technical field, some improvements and refinements without departing from the principles of the present application shall also be considered as falling within the protection scope of the present application.

Claims

1. An intelligent express delivery order processing optimization system based on big data analysis, characterized in that: include: The data acquisition module is used to collect data related to express delivery orders within the target express delivery company to obtain the corresponding logistics order information; The logistics order information includes express delivery order information and order auxiliary data; The data processing module is used to process the collected logistics order information and obtain the corresponding order decision tree diagram. The order processing module is used to obtain orders to be processed and, in conjunction with the order decision tree diagram, formulate corresponding initial order processing strategies for the respective orders to be processed. The strategy verification module is used to verify the obtained initial order processing strategy and take corresponding feedback measures based on the strategy verification results. The process of processing the collected logistics order information includes: The collected logistics order data is preprocessed to obtain the corresponding initial logistics information; the data preprocessing includes data cleaning and data integration. Obtain order data from the initial logistics information, and filter the corresponding order data to obtain the corresponding order indicator data; The corresponding order indicator data is classified according to a pre-constructed indicator classification function to obtain the corresponding order feature data; the order feature data includes a first order feature and a second order feature. Based on the initial logistics information, the time consumed by the corresponding order data processing process is obtained and marked as the order processing time, and a mapping relationship between the order processing time and the order feature data is constructed. Obtain the flowchart of the order processing, map each feature indicator in the order feature data to the corresponding order processing flowchart, and obtain the correlation between different feature indicators based on it; the feature indicator is a first order feature or a second order feature; The frequency of each feature indicator in the order data within different initial logistics information is obtained, and the frequency feature value corresponding to the corresponding feature indicator is obtained based on it. Sort the frequency feature values ​​of the first order feature and the second order feature respectively, and starting from the root node, add the corresponding first order feature and second order feature to the tree according to the sorting of the frequency feature values ​​to obtain the corresponding first-level indicator branch and second-level indicator branch. Based on the correlation between the obtained feature indicators, the corresponding first-level indicator branches and second-level indicator branches are merged to obtain the corresponding order branch tree.

2. The intelligent express delivery order processing optimization system based on big data analysis according to claim 1, characterized in that, The process of collecting data related to express delivery orders within the target courier company to obtain the corresponding logistics order information includes: The data acquisition module is equipped with several data acquisition nodes. Based on the data acquisition nodes, it collects data related to express delivery orders stored in the management center of the target express delivery company to obtain several express delivery order information, including user data, order data and delivery data. The data acquisition node is also used to collect data from external data sources to obtain corresponding order auxiliary data; the external data sources include meteorological bureaus and social media. The corresponding express delivery order information and the corresponding order auxiliary data during the order processing are summarized to obtain the corresponding logistics order information.

3. The intelligent express delivery order processing optimization system based on big data analysis according to claim 1, characterized in that, The process of classifying corresponding order indicator data based on a pre-built indicator classification function to obtain the corresponding order feature data includes: Define the index classification function as follows: In the formula, This indicates normalization processing; Indicates the first Within the initial order data, the first The regularization coefficient corresponding to each order metric data. ; Indicates the first Regression coefficients for individual order metrics data; Represents the regularization factor; Indicates the first Within the initial order data, the first Individual order metrics data; Indicates the first Reference values ​​for individual order metrics; and It is an integer. This indicates the total number of order data; An index representing order data; >0 and It is an integer. This indicates the total number of order indicator data within the corresponding order data; This represents an index for order indicator data; Obtain the regularization coefficients corresponding to each order indicator data within the relevant order data, and compare them with the pre-set regularization threshold range; If the regularization coefficient is greater than the maximum value of the regularization threshold range, the corresponding order indicator data is marked as the first order feature; if the regularization coefficient is within the regularization threshold range, the corresponding order indicator data is marked as the second order feature; if the regularization coefficient is less than the minimum value of the regularization threshold range, no other operation is performed. The corresponding first order features and second order features are summarized to obtain the corresponding first order data; Obtain the first order data corresponding to the order data in all initial logistics information, and merge the data to obtain the corresponding order feature data.

4. The intelligent express delivery order processing optimization system based on big data analysis according to claim 1, characterized in that, The process of obtaining the order decision tree diagram includes: Based on the process of obtaining the order branch tree, the corresponding user branch tree and delivery branch tree are obtained respectively; the mapping relationship between order data, delivery data and user data and the corresponding order processing time is obtained respectively; the corresponding user branch tree, delivery branch tree and order branch tree are merged to obtain the corresponding order decision tree diagram.

5. The intelligent express delivery order processing optimization system based on big data analysis according to claim 4, characterized in that, The process of acquiring pending orders and formulating corresponding initial order processing strategies for each pending order based on the order decision tree diagram includes: Based on the data acquisition node, the pending order information corresponding to the constructed pending orders is obtained from the management platform; the pending order information includes order data and user data; Based on the order information to be processed, traverse the corresponding order decision tree diagram and perform node similarity matching to obtain the root node in the corresponding order information to be processed that is similar to the corresponding order data and user data. Based on the mapping relationship between the corresponding root node and the order processing time, the order processing time of the corresponding pending order is predicted to obtain the corresponding expected order processing time. The obtained order information and expected order processing time are input into the pre-built order decision model to obtain the corresponding initial order processing strategy.

6. The intelligent express delivery order processing optimization system based on big data analysis according to claim 5, characterized in that, The process of constructing the order decision model includes: Obtain the constructed order decision tree diagram, and construct several training data sets based on it, wherein the training data sets consist of several training subsets; The network framework of the order decision model is defined as a convolutional neural network; The basic architecture of the convolutional neural network consists of an input layer, a hidden layer, and an output layer; the input layer is used to receive the input training subset and perform a linear transformation on it to obtain the corresponding feature matrix; The hidden layer is used to perform nonlinear transformations on the corresponding feature matrix, extract the features of the feature matrix, learn the patterns and rules in the data, and pass them to the output layer for decision-making. The output layer is used to receive the output of the hidden layer and make decisions on the input data based on the features extracted by the hidden layer and the patterns and rules in the learning data. Define the objective function of the order decision model. In the formula, express Real-time order processing strategy; Statement of expected value The maximum value; express Discount factor of time; and These represent the policy state and policy action at time t, respectively. Represents the reward function; ; This indicates the total expected processing time for the entered order; Define the loss function of the order decision model as follows: In the formula, Indicates the input number of the first... Labels for a training subset and It is an integer. This represents the total number of training subsets; Indicates the first Each sample corresponds to the expected output result; the corresponding order decision model is iteratively trained based on the training dataset, and based on... The optimizer continuously optimizes the model parameters of the corresponding order decision model during the iterative training process until the difference between the loss function and the corresponding objective function within the order decision model is continuous. If the training subset no longer decreases or changes, the model parameters are saved, thus completing the process of building the order decision model; among which, It is a constant.

7. The intelligent express delivery order processing optimization system based on big data analysis according to claim 5, characterized in that, The process of validating the initial order processing strategy includes: Obtain the entire process of order processing and build a corresponding order simulation model based on it; The initial order processing strategy and real-time collected order auxiliary data are input into the corresponding order simulation model, and the running process of the corresponding order simulation model is recorded to obtain the corresponding simulation order information. Obtain the complete process of order processing and divide it into detailed steps; obtain several order processing sub-processes; Based on the simulated order information, obtain the process data corresponding to different order sub-processes and mark them as first-level parameters; mark the sub-data within the corresponding process data as second-level parameters; Construct a fuzzy process evaluation interval, which contains several sub-intervals; Obtain the membership degree of the corresponding secondary parameter data and the corresponding fuzzy process evaluation interval; Obtain the membership degree between the corresponding secondary parameter data and all sub-intervals, and select the two sub-intervals with the highest membership degree respectively. and And the membership degree corresponding to the corresponding sub-interval is denoted as and ; Obtain the corresponding secondary parameters and corresponding sub-intervals The associated parameters between them; Obtain the correlation parameters between all secondary indicators within the corresponding primary indicator and the corresponding fuzzy process evaluation interval, and construct the corresponding secondary correlation matrix based on them; Obtain the product of the weight row vectors corresponding to each secondary parameter and the corresponding correlation matrix, obtain the correlation factors between the corresponding primary parameter and the corresponding fuzzy level, and construct the primary correlation matrix based on them; Obtain the product between the weight row vector corresponding to each first-level parameter and the corresponding first-level correlation matrix to obtain the strategy feature value corresponding to the initial order strategy. Set feature thresholds and compare the obtained policy feature values ​​with the corresponding feature thresholds; If the feature value of the strategy is not less than the feature threshold, then the corresponding initial order processing strategy is taken as the optimal order processing strategy, and corresponding process execution instructions are generated based on it, and instruction feedback is performed. If the strategy feature value is less than the feature threshold, then the sub-intervals corresponding to each secondary parameter are obtained, and the upper limit of the corresponding sub-interval is determined based on whether it is lower than the preset parameter threshold. If it is not lower, no other operation is performed; if it is lower, the corresponding secondary parameter is marked as a parameter to be optimized, and the corresponding parameter to be optimized is determined based on the order decision tree diagram to determine whether it is a feature indicator. If it is a feature indicator, the corresponding indicator data to be optimized is used as a constraint condition and input into the corresponding order decision model to obtain a new initial order processing strategy, and so on; if it is not a feature indicator, the corresponding indicator to be optimized is fed back to the management platform for manual intervention.

8. The intelligent express delivery order processing optimization system based on big data analysis according to claim 7, characterized in that, The membership degree between the secondary parameter data and the corresponding fuzzy process evaluation interval In the formula, Indicates the first Within the first-level parameter, the first The second-level parameter and the fuzzy process evaluation interval within the first... Membership degree between sub-intervals; and These represent the 1st, 2nd, and 3rd digits within the fuzzy process evaluation interval, respectively. The upper and lower limits of each sub-interval; Indicates the first Within the first-level parameter, the first There are two secondary parameters; among them, and It is an integer. Indicates the total number of sub-processes in the order; and It is an integer. This indicates the total number of secondary parameters within the corresponding order sub-process; Corresponding secondary parameters and corresponding sub-intervals The correlation parameters between In the formula, Indicates the first Within the first-level parameter, the first Within the evaluation range of each secondary parameter and the fuzzy process The correlation parameters between sub-intervals.

9. A method for optimizing intelligent express delivery order processing based on big data analysis, implemented based on the intelligent express delivery order processing optimization system based on big data analysis as described in any one of claims 1 to 8, characterized in that, include: Step 1: Collect data related to express delivery orders from the target courier company to obtain the corresponding logistics order information; The logistics order information includes express delivery order information and order auxiliary data; Step 2: Process the collected logistics order information to obtain the corresponding order decision tree diagram; Step 3: Obtain the orders to be processed and formulate the corresponding initial order processing strategies for the orders based on the order decision tree diagram; Step 4: Validate the obtained initial order processing strategy and take corresponding feedback measures based on the strategy validation results.

Images