A logistics offer management method and system

Abstract

The application discloses a kind of logistics offer management method and system, belong to logistics offer management technical field, its method specifically includes: the historical behavior data of user logistics offer is collected, user behavior model is established based on the historical behavior data, the user behavior model is used to characterize the game behavior of user, according to user behavior model, the behavior mode that user influences initial offer strategy is identified, corresponding disturbance factor is generated, initial offer strategy is adjusted based on disturbance factor, obtains the offer result containing user behavior disturbance, according to the deviation between order fulfillment and user actual behavior, dynamically update user behavior model and disturbance factor;The application can realize active identification and inhibition to user behavior disturbance in the process of offer generation, make offer calculation have self-adaptability and stability, improve processing consistency and response efficiency under complex interactive scene.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of logistics quotation management, and in particular to a logistics quotation management method and system. BACKGROUND

[0002] Logistics quotation management is an important part of a logistics information system, and its main function is to provide corresponding transportation price schemes for users according to transportation paths, storage locations, cargo characteristics, delivery timeliness and various other factors. Existing logistics quotation management systems are generally based on historical transportation costs, market price rules and storage inventory conditions, and combine preset calculation models to output quotation results.

[0003] In actual applications, the logistics quotation process not only depends on data calculation within the system, but also is affected by the interactive behaviors of users in the quotation stage. For example, some users frequently initiate quotation requests, change cargo sorting strategies, select specific time windows to place orders, or even use repeated trial-and-error methods to evade the default quotation path of the system within a short period of time. These behaviors may cause price fluctuations or strategy imbalances in the system during operation.

[0004] Existing logistics quotation management methods usually correct quotations by setting fixed rules, simply adjusting parameter weights, or introducing compensation factors in the later calculation stage. However, such methods are difficult to accurately identify and predict complex behavior patterns of users, and lack dynamic adjustment mechanisms for behavior disturbances. When a game relationship is formed between the quotation strategy and the user behavior, stability and calculation consistency are easily affected, which may cause unnecessary multiple rounds of recalculation, increase the occupation of computing resources and prolong the response time. SUMMARY

[0005] To overcome the deficiencies of the prior art, the application provides a logistics quotation management method and system.

[0006] To achieve the above purpose, the application provides the following technical solutions:

[0007] A logistics quotation management method, comprising:

[0008] Collecting historical behavior data of user logistics quotation, establishing a user behavior model based on the historical behavior data, the user behavior model being used to represent the game behavior of the user;

[0009] According to the user behavior model, identifying the behavior pattern of the user affecting the initial quotation strategy, and generating a corresponding disturbance factor;

[0010] Adjusting the initial quotation strategy based on the disturbance factor to obtain a quotation result containing user behavior disturbance;

[0011] According to the deviation between the order fulfillment and the actual behavior of the user, the user behavior model and the disturbance factor are dynamically updated.

[0012] Specifically, the historical behavior data of the user logistics quotation is collected, and a user behavior model is established based on the historical behavior data, including:

[0013] The historical behavior data of the user logistics quotation includes refresh frequency, order splitting operation, order placement time selection, and path preference.

[0014] Identify the target user identifier, associate the historical order record, the quotation request record and the corresponding user account, and construct a behavior tracking sequence.

[0015] The operations in the behavior tracking sequence are semantically classified and labeled as preset behavior labels, including browsing operation, comparison operation, avoidance operation, repeated operation and time point decision operation.

[0016] Based on the time sequence distribution characteristics between operation labels, the behavior transition mode is extracted, and the user behavior model is constructed based on operation nodes.

[0017] Specifically, according to the user behavior model, the behavior mode of the user affecting the initial quotation strategy is identified, and the corresponding disturbance factor is generated, including:

[0018] Extract the decision node sequence of the target user in the user behavior model, and cluster the operation jump mode between adjacent nodes.

[0019] Backtrack the node combination in each cluster in reverse order, identify the behavior segment of the user actively skipping the recommended strategy or repeatedly exploring the path in the quotation process.

[0020] Based on the cross-overlapping degree of the behavior segment between users, a similarity matrix of intervention behavior is constructed.

[0021] Map the intervention behavior similarity matrix to a multi-dimensional behavior intervention space, generate a disturbance factor for representing the disturbance tendency of user behavior by positioning the dense trajectory cluster in the frequent disturbance area.

[0022] Specifically, the similarity matrix of intervention behavior is constructed based on the cross-overlapping degree of the behavior segment between users, including:

[0023] Each behavior segment is encoded into a segment identifier string according to the node type and operation sequence.

[0024] The segment identifier string is searched in the behavior sequence of different users, and the user identifier set where it appears is recorded.

[0025] Calculate the user identifier set intersection of any two segment identifier strings to form a segment association pair.

[0026] Based on the intersection size of the fragment association pairs and the difference in the operation sequence, the corresponding similarity values ​​are generated and filled into the corresponding positions of the intervention behavior similarity matrix to construct the intervention behavior similarity matrix.

[0027] Specifically, the intervention behavior similarity matrix is ​​mapped to a multi-dimensional behavior intervention space. By locating dense trajectory clusters in frequently disturbed regions, a perturbation factor representing the user's behavioral perturbation tendency is generated, including:

[0028] Assign corresponding multidimensional spatial coordinates to each behavior segment in the intervention behavior similarity matrix, with the coordinate dimensions corresponding to the preset intervention feature categories;

[0029] The spatial coordinates of all behavioral fragments are combined to form a behavioral intervention point set, and a neighborhood search is performed in the point set to determine candidate trajectory clusters.

[0030] Sequence order analysis is performed on the point set within each candidate trajectory cluster to filter out the trajectory set with cross-node jump characteristics;

[0031] Based on the density of node distribution and the frequency of jumps in the trajectory set, a corresponding perturbation factor is generated.

[0032] Specifically, the step of adjusting the initial pricing strategy based on the disturbance factor to obtain a pricing result that includes user behavior disturbances includes:

[0033] Identify the behavioral category labels in the disturbance factors and map them to a preset set of price adjustment parameters;

[0034] In the quotation calculation process, the set of quotation adjustment parameters is inserted during the parameter initialization stage to replace or correct the original parameter values.

[0035] In the intermediate stage of the quotation calculation, for calculation nodes involving path selection, warehouse allocation and time window matching, the parameter branch corresponding to the disturbance factor is called.

[0036] Before generating the quotation results, a global consistency check based on the perturbation factor is performed on the output of all computing nodes, and the necessary quotation recalculation process is retried based on the check results.

[0037] Specifically, in the intermediate stage of price calculation, for calculation nodes involving path selection, warehouse allocation, and time window matching, the parameter branch corresponding to the disturbance factor is invoked, including:

[0038] Determine the set of calculation nodes that fall within the path selection, warehouse allocation, and time window matching range in the current quotation calculation process;

[0039] retrieving a perturbation factor type label associated with each node in the set of computing nodes;

[0040] loading a corresponding parameter branch configuration from a preset parameter branch library according to the perturbation factor type label;

[0041] injecting the parameter branch configuration into a corresponding computing node operation flow to replace an original parameter set of the node;

[0042] completing subsequent computation of the computing node after the parameter replacement according to a preset execution sequence.

[0043] Specifically, the dynamic updating of the user behavior model and the perturbation factor according to the deviation between the order fulfillment and the actual behavior of the user includes:

[0044] extracting fulfillment path data, delivery timing data, and warehouse calling data corresponding to the target order from the fulfillment record;

[0045] aligning the fulfillment path data with the actual behavior sequence of the user in the whole process from ordering to fulfillment completion to generate a behavior-fulfillment mapping table;

[0046] calculating a deviation index between a fulfillment execution node and a corresponding user operation node in the behavior-fulfillment mapping table;

[0047] mapping the deviation index to a corresponding feature parameter of the user behavior model, and dynamically updating the user behavior model and the perturbation factor.

[0048] A logistics pricing management system for implementing the logistics pricing management method, comprising a model generation module, a perturbation module, a pricing module, and a dynamic updating module.

[0049] The model generation module is configured to collect historical behavior data of user logistics pricing, and establish a user behavior model based on the historical behavior data.

[0050] The perturbation module is configured to identify a behavior pattern of the user affecting an initial pricing strategy according to the user behavior model, and generate a corresponding perturbation factor.

[0051] The pricing module is configured to adjust the initial pricing strategy based on the perturbation factor to obtain a pricing result containing user behavior perturbation.

[0052] The dynamic updating module is configured to dynamically update the user behavior model and the perturbation factor according to the deviation between the order fulfillment and the actual behavior of the user.

[0053] Specifically, the pricing module comprises a mapping unit, an adjustment unit, and a verification unit.

[0054] The mapping unit is configured to identify the behavior category label in the disturbance factor and map it to a preset offer adjustment parameter set.

[0055] The adjustment unit is configured to adjust the offer parameters in the offer calculation stage.

[0056] The verification unit is configured to perform a global consistency check based on the disturbance factor on the outputs of all the calculation nodes before generating the offer result, and retrigger the necessary offer recalculation process according to the check result.

[0057] Compared with the prior art, the present application has the following advantages:

[0058] The present application provides a logistics offer management method and system, which collects multi-dimensional historical behavior data of users in the logistics offer process, constructs a user behavior model with game attributes, predicts the disturbance behavior of users on the offer strategy based on the model, generates corresponding disturbance factors, dynamically introduces the disturbance factors into key calculation nodes such as path selection, warehouse allocation and time window matching in the middle link of offer calculation, outputs the offer result, and updates the user behavior model and disturbance factors in a closed loop according to the deviation between the actual behavior of the user and the fulfillment data after the order is fulfilled. The method can actively identify and suppress the disturbance of user behavior during the offer generation process, making the offer calculation adaptive and stable, thereby improving the processing consistency and response efficiency in complex interactive scenarios, and reducing the burden of repeated calculation and resource occupation. BRIEF DESCRIPTION OF DRAWINGS

[0059] Figure 1 The flowchart provided by the present application;

[0060] Figure 2 The system architecture diagram provided by the present application. DETAILED DESCRIPTION

[0061] The present application will be described in detail below with reference to specific embodiments. The following embodiments will help those skilled in the art to further understand the present application, but do not limit the present application in any form. It should be noted that those skilled in the art can make several modifications and improvements without departing from the concept of the present application. These all belong to the protection scope of the present application.

[0062] In order to make the purpose, technical scheme and advantages of the present application clearer and more understandable, the present application will be further described in detail below with reference to the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and do not limit the present application.

[0063] It should be noted that the various features of the embodiments of the present application can be combined with each other, and all are within the protection scope of the present application, if there is no conflict. In addition, although the functional modules are divided in the device schematic diagram, and the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in a different order from the module division in the device or the order in the flowchart. In addition, the "first", "second", "third", etc. used in the present application do not limit the data and execution order, and only distinguish the same items or similar items with basically the same function and effect.

[0064] The terms "first", "second", "third", etc. do not limit the data and execution order, and only distinguish the same items or similar items with basically the same function and effect.

[0065] Unless otherwise defined, all technical and scientific terms used in the present application have the same meaning as understood by a person skilled in the art to which the present application belongs. The terms used in the present application are only for the purpose of describing the specific embodiments of the present application, and are not used to limit the present application. The term "and / or" used in the present application includes any and all combinations of one or more related listed items.

[0066] Embodiment 1

[0067] Please refer to Figure 1 The present application provides an embodiment: a logistics quotation management method, comprising the following specific steps:

[0068] Step S1: collecting historical behavior data of user logistics quotation, establishing a user behavior model based on the historical behavior data, the user behavior model is used to represent the user's game behavior.

[0069] The specific steps of step S1 are:

[0070] Step S101: collecting historical behavior data of user logistics quotation, including refresh frequency, order splitting operation, order time selection, path preference.

[0071] Step S102: identifying target user identification, associating historical order records, quotation request records and corresponding user accounts to build a behavior tracking sequence.

[0072] In the embodiment, according to the unique account identifier of the target user on the quotation platform, the historical order record and the historical quotation request record of the user are called from the data storage module. The historical order record includes order number, cargo type, starting warehouse, destination, transportation path identifier, and order time, etc. The quotation request record includes quotation request time, selected parameter configuration, and quotation scheme identifier returned by the system, etc. Then, the above two types of records are uniformly sorted according to the time axis, and are associated based on the account identifier to generate a behavior event sequence arranged in time sequence. Finally, each record in the event sequence is labeled with a corresponding behavior type label, such as quotation query, order confirmation, order modification, etc.

[0073] Step S103: Semantically classifying the operations in the behavior tracking sequence and marking as preset behavior labels, including browsing operation, comparison operation, avoidance operation, repeated operation, and time point decision operation.

[0074] In the embodiment, first, all event records in the behavior tracking sequence generated in step S102 are read, and the operation type, trigger condition, and context parameter of each event are extracted. Then, the semantic classification module is called to match the extracted event attributes with the preset behavior label classification rules. The label rules are a classification mapping table established by analyzing historical user behavior characteristics, including browsing operation, comparison operation, avoidance operation, repeated operation, and time point decision operation, etc. For example, if consecutive quotation query events repeatedly appear under the condition that the parameter change amplitude is small, the operation is marked as comparison operation. If orders are submitted at a specific time point to avoid time period additional fees, the operation is marked as time point decision operation. After the matching is completed, each event is assigned with the corresponding label, and the original event type field in the behavior tracking sequence is replaced with the label to obtain a behavior sequence with semantic annotation.

[0075] Step S104: Based on the time sequence distribution characteristics between operation labels, behavior transition patterns are extracted, and a user behavior model is constructed based on operation nodes.

[0076] In the embodiment, the behavior tracking sequence with semantic labels marked in step S103 is read, and the distribution of each operation label on the time axis is analyzed, including the appearance order of adjacent labels, the appearance interval, and the position characteristics in the sequence. Then, the jump relationship between each label is taken as a candidate behavior transition unit, and these units are classified and counted according to the label combination type to obtain the transition probability distribution between labels and the common jump path set. On this basis, each label is regarded as an operation node in the user behavior model, and the jump relationship between different labels is regarded as a directed connection edge between nodes to form a user behavior model based on operation nodes.

[0077] Step S2: According to the user behavior model, identify the behavior pattern of the user affecting the initial offer strategy, and generate the corresponding disturbance factor.

[0078] The specific steps of step S2 are:

[0079] Step S201: Extract the decision node sequence of the target user in the user behavior model, and cluster the operation jump patterns between adjacent nodes.

[0080] In this embodiment, all operation nodes corresponding to the target user identification are retrieved from the constructed user behavior model, and the decision node sequence is extracted according to the time sequence connection relationship of the nodes in the model; the decision node sequence reflects the user's behavior path in a specific offer process; then, the operation difference characteristics between adjacent nodes are calculated, including label category change, time interval between nodes, and strategy branch identification of the previous and next nodes; then, these adjacent node pairs are regarded as the basic unit of jump pattern, the feature encoding of all jump patterns is performed, and the clustering analysis is performed based on similarity measurement; finally, the clustering result obtained is used to represent the operation switching mode distribution of the user in different behavior stages.

[0081] Step S202: Backtrack the node combination in each cluster in reverse order to identify the behavior segment of the user actively skipping the recommended strategy or repeatedly exploring the path in the offer process.

[0082] In this embodiment, for each jump pattern cluster obtained in step S201, the node combination in each cluster is traversed in turn, and the time sequence order of the node combination is reversed to form a reverse order node chain; then, the connection relationship between the nodes and the associated offer strategy identification are analyzed by backtracking from the end node to the front along the reverse order chain; when it is detected in the backtracking process that there is a jump around the system recommended path in the node combination, or there is a loop switching that repeatedly goes back and forth between the same path nodes, the part of the node chain is cut off as a behavior segment.

[0083] Step S203: Construct an intervention behavior similarity matrix based on the cross-overlapping degree of the behavior segments between users.

[0084] The specific steps of step S203 are:

[0085] Step S2031: Encode each behavior segment into a segment identification string according to the node type and operation order.

[0086] Step S2032: Retrieve the segment identification string in the behavior sequence of different users, and record the user identification set in which it appears.

[0087] In the embodiment, the behavior segment identification string generated by the previous processing step is received, and the identification string is combined by the label sequence of each operation node in the segment and the connection order thereof; subsequently, the multi-user historical behavior sequence stored in the behavior dataset is traversed, and a sliding window search is performed on each behavior sequence in chronological order, and whether the operation label combination in the window is completely identical to the target identification string is compared; if the matching is successful, the unique identification of the user corresponding to the behavior sequence is read, and the unique identification is added to the user identification set corresponding to the segment identification string; after the search is completed, a mapping table of the user identification set and the occurrence number is established for each segment identification string.

[0088] Step S2033: calculating the user identification set intersection of any two segment identification strings to form a segment association pair.

[0089] In the embodiment, the user identification set corresponding to each segment identification string is read from the result of step S2032; subsequently, in the set list of all segment identification strings, any two different segment identification strings are sequentially selected in a combined manner, and the set intersection operation is performed on the user identification sets of the two segment identification strings to obtain a user identification list that is simultaneously contained in the two sets; then, the user identification list is recorded together with the corresponding two segment identification strings to form a segment association pair, and the segment association pair is assigned a unique association pair number; finally, all generated segment association pairs are stored in the segment association relationship table.

[0090] Step S2034: generating a corresponding similarity value according to the intersection size of the segment association pair and the operation sequence difference, and filling the similarity value into the corresponding position of the intervention behavior similarity matrix to construct the intervention behavior similarity matrix.

[0091] In the embodiment, for each segment association pair generated in step S2033, the intersection size of the corresponding user identification set and the operation sequence information of the two segment identification strings in the association pair are read; subsequently, the difference degree index between the two operation sequences is calculated, and the difference degree is composed of the operation node type difference and the node order difference; then, the intersection size and the difference degree are converted into a similarity value according to a preset numerical mapping rule, and the value range of the similarity value corresponds to the matrix index; finally, the similarity value is filled into the intervention behavior similarity matrix with the two segment identification strings as the row and column indices, until the similarity of all segment association pairs is filled, thereby constructing a complete intervention behavior similarity matrix.

[0092] Step S204: mapping the intervention behavior similarity matrix to a multi-dimensional behavior intervention space, and generating a disturbance factor for representing the user behavior disturbance tendency by positioning the dense trajectory cluster in the frequently disturbed area.

[0093] The specific steps of step S204 are as follows:

[0094] Step S2041: Assign a corresponding multi-dimensional space coordinate to each behavior segment in the intervention behavior similarity matrix, and the coordinate dimension corresponds to the preset intervention feature category.

[0095] In this embodiment, all behavior segment identifiers in the intervention behavior similarity matrix are read, and the intervention feature data related to each segment is retrieved, including but not limited to time sensitivity, path deviation degree, node repetition rate, strategy avoidance strength, and cross-stage switching frequency, etc. Then, a coordinate dimension is assigned to each intervention feature category, and the feature value of the segment is mapped to the numerical coordinate position of the dimension according to the preset numerical standard. Then, the coordinate values of all dimensions are combined into the multi-dimensional space coordinate of the behavior segment, and stored in the intervention feature space data table.

[0096] Step S2042: Combine the space coordinates of all behavior segments to form a behavior intervention point set, and perform neighborhood search in the point set to determine the candidate trajectory cluster.

[0097] In this embodiment, all behavior segment multi-dimensional space coordinates generated in step S2041 are read in sequence, and combined into a behavior intervention point set according to the space coordinate values, each point in the point set corresponding to a multi-dimensional feature representation of a behavior segment. Then, based on the preset spatial neighborhood radius parameter, the neighborhood search operation is performed on the point set to identify the point set that is close to each other in space. In the neighborhood search process, the spatial distance of each point from the surrounding points is calculated, and the number of neighbor points with a distance not exceeding the radius is counted. If the number of neighbors of a certain segment point exceeds the preset threshold, the point and its neighboring points in the neighborhood are classified into the same candidate trajectory cluster. After the traversal is completed, the data set of all candidate trajectory clusters is output.

[0098] Step S2043: Perform sequence order analysis on the point set in each candidate trajectory cluster to filter out the trajectory set with the cross-node jumping feature.

[0099] In this embodiment, for each candidate trajectory cluster obtained in step S2042, the original behavior segment sequence data corresponding to all points in the cluster is read. Then, the segments in the same trajectory cluster are sorted according to their occurrence order in the original user behavior sequence, and the jump relationship and direction between nodes are recorded. Then, cross-node jumping detection is performed on the sorted sequence, which specifically identifies the operation mode of directly jumping to a non-adjacent node without passing through the consecutive adjacent nodes in the model. If multiple segments in the same trajectory cluster are found to have similar cross-node jumping patterns, these segments are integrated into a trajectory set, and the set is assigned a unique identifier.

[0100] Step S2044: generating a corresponding perturbation factor based on the density of node distribution in the trajectory set and the frequency of cross-jump.

[0101] In this embodiment, for each trajectory set obtained in step S2043, the distribution density of each node in the user behavior model is counted, which is determined by the number of times the node appears in the trajectory set and the proportion of the total number of reachable paths in the model. Then, the frequency of cross-node jump events in the trajectory set is calculated, which is the ratio of the number of jump events to the total number of events in the trajectory set. Finally, the node distribution density and the frequency of cross-jump are input into the perturbation factor generation module, and the perturbation factor value is converted according to the preset mapping rule, and a unique perturbation factor identifier is generated for each trajectory set.

[0102] Step S3: adjusting the initial offer strategy based on the perturbation factor to obtain an offer result containing user behavior perturbation.

[0103] The specific steps of step S3 are:

[0104] Step S301: identifying the behavior category label in the perturbation factor and mapping it to a preset offer adjustment parameter set.

[0105] In this embodiment, the perturbation factor data structure generated in the previous step is read to parse the behavior category label associated with the factor, which includes but is not limited to price exploration, path avoidance, time window concentration, repeated offer, and warehouse switching. Then, the offer adjustment parameter mapping table stored in the parameter configuration library is accessed, in which each type of behavior label corresponds to a set of adjustment parameters that can be used for offer calculation. The adjustment parameter set may include basic freight correction coefficient, path weight adjustment value, time window priority parameter, and inventory call threshold. Then, the behavior category label parsed is matched with the entries in the mapping table to extract the corresponding offer adjustment parameter set, and the set is bound with the unique identifier of the perturbation factor.

[0106] Step S302: inserting the offer adjustment parameter set in the parameter initialization stage in the offer calculation process to replace or modify the original parameter value.

[0107] In the embodiment, when the quotation engine starts the quotation calculation process, it is positioned to the parameter initialization stage, which is used to load initial parameters for each calculation module, including basic freight, path cost weight, warehouse calling priority, time window additional coefficient, etc. Then, all parameter items are read from the quotation adjustment parameter set bound in step S301, and compared with the original parameters loaded in the initialization stage item by item. When it is detected that the corresponding parameter item already exists, the value in the adjustment parameter set is directly used to replace the original parameter value. If the corresponding parameter item does not exist, the adjustment parameter is inserted into the parameter list as a new item. After the replacement and insertion operations are completed, the parameter configuration file of the initialization stage is regenerated, and is passed to the subsequent quotation calculation.

[0108] Step S303: In the middle stage of the quotation calculation, the parameter branch corresponding to the perturbation factor is called for the calculation nodes related to path selection, warehouse allocation and time window matching.

[0109] The specific steps of step S303 are as follows:

[0110] Step S3031: Determine the calculation node set in the current quotation calculation process that is within the range of path selection, warehouse allocation and time window matching.

[0111] In the embodiment, the execution plan table of the current task is read in the calculation link of the quotation engine. The plan table lists all the calculation nodes participating in the quotation calculation and their function types in module order. Then, the nodes whose function types are path selection, warehouse allocation and time window matching are filtered out from the nodes according to the function identification of the nodes, and the node reference addresses of the nodes meeting the conditions are stored in a node set structure. For the case of a composite function node, for example, a node related to path selection and warehouse allocation, multiple labels are marked in the set, so that the corresponding function logic can be called respectively in subsequent processing. Finally, the generated calculation node set is input as the target set of the subsequent parameter calling and strategy adjustment steps.

[0112] Step S3032: For each node in the calculation node set, the perturbation factor type label associated with the node is retrieved.

[0113] In the embodiment, the calculation node set generated in step S3031 is read, and each node in the set is traversed in turn. For each node, its node identification information and function type identification are obtained, and the perturbation factor index table is accessed. Then, in the perturbation factor index table, the node identification information is used as the retrieval key to query the perturbation factor record associated with the node, and the perturbation factor type label in the record is parsed. If a node is associated with multiple perturbation factors, all the type labels are stored in the associated label field of the node in the form of a list. After the traversal is completed, a mapping table of node identification-perturbation factor type label set is formed.

[0114] Step S3033: According to the disturbance factor type label, load the corresponding parameter branch configuration from the preset parameter branch library.

[0115] In this embodiment, the node identification-disturbance factor type label set mapping table generated in step S3032 is read, and each node disturbance factor type label is traversed in turn. For each type label, the preset parameter branch library is accessed, which is partitioned and managed according to the behavior intervention category, and each partition stores a plurality of parameter branch configuration files. Then, the corresponding parameter branch configuration file is retrieved in the partition matched with the label, and all parameter items of the configuration file are loaded into the temporary buffer area. If a node is associated with multiple disturbance factor type labels, the corresponding parameter branches are loaded in the order of the labels, and the parameter items are merged or overwritten in the buffer area. After the loading is completed, the parameter branch configuration of the node is cached to the node execution context.

[0116] Step S3034: Inject the parameter branch configuration into the corresponding calculation node operation flow to replace the original parameter set of the node.

[0117] In this embodiment, the parameter branch configuration loaded in step S3033 is read, and the calculation node bound thereto is located. Then, before the operation flow of the calculation node is started, the original parameter set of the node is copied to a temporary buffer area for backtracking use. Then, the original parameter set of the node is replaced item by item using the parameter values in the parameter branch configuration. For a field that does not exist in the original parameter set, it is directly appended to the parameter structure and registered as a valid calculation variable in the execution context of the node. After the parameter replacement and appending operations are completed, the updated parameter set is written back to the running environment of the node.

[0118] Step S3035: For the calculation node after the parameter replacement, the subsequent calculation is completed according to the preset execution order.

[0119] In this embodiment, for the calculation node after the parameter replacement in step S3034, the order index value of the node in the execution plan table is read, and the subsequent node dependency relationship of the node in the calculation link is determined. Then, according to the preset execution order control table, the calculation processes of the dependent nodes are triggered in turn starting from the current node, wherein the current node first executes the operation logic such as path selection, warehouse allocation or time window matching according to the injected new parameter set, and writes the result data to the intermediate result buffer area. Then, the downstream nodes dependent on the output of the node acquire the intermediate results in order and continue to execute their corresponding calculation operations, until all the nodes in the branch link complete the calculation tasks. Finally, the result data of the calculation link is submitted to the global aggregation.

[0120] Step S304: Before generating the quotation result, the outputs of all computing nodes are checked for global consistency based on the perturbation factors, and the necessary quotation recalculation process is retriggered according to the checking result.

[0121] In this embodiment, before the quotation calculation process enters the result aggregation stage, all computing node outputs in the global output data buffer are called, and their positions in the link and dependency relationships are determined according to the node execution plan table; then, all perturbation factors and their type labels that have participated in the calculation are read from the perturbation factor index table, and they are compared with the outputs of the corresponding computing nodes one by one. For the nodes associated with the perturbation factors, consistency checking logic is performed, including but not limited to: value range checking of key fields before and after parameter adjustment, cross-node data format matching verification, and link value balance checking. If the checking result shows that there is a deviation exceeding the threshold or a dependency conflict, the node and its downstream link are marked as needing to be recalculated in the execution control module; then, the corresponding quotation recalculation process is retriggered according to the marked state, and only the affected node link is recalculated, and the new result data is updated to the global output buffer after completion.

[0122] Step S4: Dynamically updating the user behavior model and the perturbation factor according to the deviation between the order fulfillment and the actual behavior of the user.

[0123] The specific steps of step S4 are:

[0124] Step S401: Extracting the fulfillment path data, delivery timing data, and warehouse calling data corresponding to the target order from the fulfillment record.

[0125] Step S402: Aligning the fulfillment path data with the actual behavior sequence of the user in the whole process from order placement to fulfillment completion to generate a behavior-fulfillment mapping table.

[0126] In this embodiment, the fulfillment path data of the target order is obtained, and the path data is composed of multiple fulfillment nodes, each node containing node identification, completion timestamp, processing action type, and other information; then, the user behavior sequence associated with the order is extracted, which covers the whole process from the user initiating the order placement request to the fulfillment completion, and is arranged in chronological order; then, the fulfillment path nodes and the user behavior events are aligned one by one with the timestamp as the primary key. For records with time overlap or interval within the preset threshold, a one-to-one mapping relationship is established. For the case where the same fulfillment node is corresponding to cross-stage or multiple behaviors, the behavior set is recorded in the mapping table. Finally, the behavior-fulfillment mapping table formed contains the pairing information of the fulfillment path nodes and the corresponding user behaviors, and retains the relative order and context identification of the behavior occurrence.

[0127] Step S403: In the behavior-compliance mapping table, the deviation index between the compliance execution node and the corresponding user operation node is calculated.

[0128] In this embodiment, the behavior-compliance mapping table generated in step S402 is read, and the mapping records therein are traversed one by one. For each record, the key indicator data of the compliance execution node is extracted, including node completion time, processing time consumption, state change times, etc., and the feature data of the corresponding user operation node is extracted, including operation trigger time, continuous operation interval, behavior frequency, etc. Then, according to the preset deviation calculation rule, the difference values of the compliance node and the user operation node in the time, times, and behavior density dimensions are compared to generate a multi-dimensional deviation vector. If a compliance node corresponds to multiple user operation nodes, the user operation data is aggregated when calculating to ensure that the indicators of the compliance node are under the same comparison benchmark. Finally, the deviation index is generated.

[0129] Step S404: Map the deviation index to the corresponding feature parameters of the user behavior model, and dynamically update the user behavior model and the disturbance factor.

[0130] In this embodiment, the deviation index table generated in step S403 is read, and according to the compliance node identifier and the user operation node identifier in the deviation index record, the corresponding feature parameter position in the user behavior model is located. Then, each dimension of the deviation index vector is one-to-one corresponding to the pre-defined feature parameters in the user behavior model, for example, the time difference dimension is mapped to the response delay parameter of the model, the frequency difference dimension is mapped to the interaction stability parameter, and the behavior density dimension is mapped to the focus concentration parameter. After the mapping is completed, the parameter values in the model are corrected according to the preset dynamic updating rule. The correction methods include direct replacement of parameter values, proportional increase or decrease, or smoothing update based on historical trends. Then, the disturbance factor weights bound to these feature parameters are recalculated.

[0131] Embodiment 2

[0132] Please refer to Figure 2 Another embodiment provided by the present application is a logistics quotation management system, comprising a model generation module, a disturbance module, a quotation module, and a dynamic updating module.

[0133] The model generation module is configured to collect historical behavior data of user logistics quotations, and establish a user behavior model based on the historical behavior data.

[0134] The disturbance module is configured to identify, according to the user behavior model, a behavior pattern of the user that has an impact on an initial quotation strategy, and generate a corresponding disturbance factor.

[0135] The offer module is configured to adjust the initial offer strategy based on the disturbance factor to obtain an offer result containing user behavior disturbance.

[0136] The dynamic updating module is configured to dynamically update the user behavior model and the disturbance factor according to the deviation between the order fulfillment and the actual user behavior.

[0137] The offer module comprises a mapping unit, an adjusting unit and a checking unit.

[0138] The mapping unit is configured to identify the behavior category label in the disturbance factor and map it to a preset offer adjustment parameter set.

[0139] The adjusting unit is configured to adjust the offer parameter in the offer calculation stage.

[0140] The checking unit is configured to perform a global consistency check based on the disturbance factor on the output of all calculation nodes before generating the offer result, and retrigger the necessary offer recalculation process according to the check result.

[0141] In addition, the part of the above technical solution in the embodiments of the present application that is consistent with the implementation principle of the corresponding technical solution in the prior art is not described in detail to avoid excessive repetition.

[0142] The specific embodiments described above further explain the purpose, technical solutions and beneficial effects of the present application. It should be understood that the above description is only a specific embodiment of the present application and is not intended to limit the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. A logistics quotation management method, characterized in that, include: Collect historical behavioral data on user logistics pricing, and build a user behavior model based on the historical behavioral data. The user behavior model is used to represent the user's game behavior. Based on the user behavior model, identify the behavioral patterns of users that affect the initial pricing strategy and generate corresponding perturbation factors; The initial pricing strategy is adjusted based on the disturbance factor to obtain pricing results that include user behavior disturbances; Based on the discrepancy between order fulfillment status and actual user behavior, dynamically update the user behavior model and disturbance factors; The step of identifying user behavior patterns that influence the initial pricing strategy based on the user behavior model and generating corresponding perturbation factors includes: Extract the decision node sequence of the target user in the user behavior model, and cluster the operation jump patterns between adjacent nodes; Reverse backtracking of node combinations within each cluster identifies user behavior segments during the pricing process where they actively skipped recommendation strategies or repeatedly explored paths. Based on the degree of overlap between the behavioral segments, an intervention behavior similarity matrix is ​​constructed; The intervention behavior similarity matrix is ​​mapped to a multi-dimensional behavior intervention space, and a perturbation factor representing the user's behavior perturbation tendency is generated by locating dense trajectory clusters in frequently disturbed areas. The process of constructing an intervention behavior similarity matrix based on the degree of overlap between the behavioral segments among users includes: Each action segment is encoded into a segment identifier string according to the node type and operation order; Retrieve the segment identifier string from the behavior sequences of different users and record the set of user identifiers in which it appears; Calculate the intersection of the user identifier sets of any two fragment identifier strings to form fragment association pairs; Based on the intersection size of the fragment association pairs and the difference in the operation sequence, the corresponding similarity values ​​are generated, and the values ​​are filled into the corresponding positions of the intervention behavior similarity matrix to construct the intervention behavior similarity matrix; The intervention behavior similarity matrix is ​​mapped to a multi-dimensional behavior intervention space. By locating dense trajectory clusters in frequently disturbed regions, a perturbation factor representing the user's behavior perturbation tendency is generated, including: Assign corresponding multidimensional spatial coordinates to each behavior segment in the intervention behavior similarity matrix, with the coordinate dimensions corresponding to the preset intervention feature categories; The spatial coordinates of all behavioral fragments are combined to form a behavioral intervention point set, and a neighborhood search is performed in the point set to determine candidate trajectory clusters. Sequence order analysis is performed on the point set within each candidate trajectory cluster to filter out the trajectory set with cross-node jump characteristics; Based on the density of node distribution and the frequency of jumps in the trajectory set, a corresponding perturbation factor is generated; The method of adjusting the initial pricing strategy based on perturbation factors to obtain pricing results that include user behavior perturbations includes: Identify the behavioral category labels in the disturbance factors and map them to a preset set of price adjustment parameters; In the quotation calculation process, the set of quotation adjustment parameters is inserted during the parameter initialization stage to replace or correct the original parameter values. In the intermediate stage of the quotation calculation, for calculation nodes involving path selection, warehouse allocation and time window matching, the parameter branch corresponding to the disturbance factor is called. Before generating the quotation results, a global consistency check based on the perturbation factor is performed on the output of all computing nodes, and the necessary quotation recalculation process is retried based on the check results.

2. The logistics quotation management method as described in claim 1, characterized in that, The process of collecting historical user behavior data on logistics pricing and establishing a user behavior model based on this historical behavior data includes: Collect historical behavioral data on user logistics pricing, including refresh frequency, order splitting operations, order placement time selection, and route preferences; Identify target user identifiers, associate historical order records and quotation request records with corresponding user accounts, and construct a behavior tracking sequence; The operations in the behavior tracking sequence are semantically classified and labeled with preset behavior tags, including browsing operations, comparison operations, avoidance operations, repeated operations, and time-point decision operations. Based on the temporal distribution characteristics between operation tags, behavior transfer patterns are extracted, and user behavior models are constructed based on operation nodes.

3. The logistics quotation management method as described in claim 2, characterized in that, In the intermediate stage of price calculation, for calculation nodes involving path selection, warehouse allocation, and time window matching, the parameter branch corresponding to the disturbance factor is invoked, including: Determine the set of calculation nodes that fall within the path selection, warehouse allocation, and time window matching range in the current quotation calculation process; For each node in the set of computing nodes, retrieve the perturbation factor type label associated with that node; Based on the disturbance factor type label, load the corresponding parameter branch configuration from the preset parameter branch library; The parameter branch configuration is injected into the corresponding computing node's operation flow, replacing the node's original parameter set; For the computation nodes with replaced parameters, subsequent calculations are completed according to the preset execution order.

4. The logistics quotation management method as described in claim 3, characterized in that, The method of dynamically updating the user behavior model and disturbance factors based on the deviation between order fulfillment status and actual user behavior includes: Extract fulfillment path data, delivery sequence data, and warehouse call data corresponding to the target order from the fulfillment records; Align the fulfillment path data with the user's actual behavior sequence throughout the entire process from order placement to fulfillment completion to generate a behavior-fulfillment mapping table; In the behavior-performance mapping table, the deviation index between the performance execution node and the corresponding user operation node is calculated; The deviation index is mapped to the corresponding feature parameters of the user behavior model, and the user behavior model and perturbation factor are dynamically updated.

5. A logistics quotation management system, used to implement the logistics quotation management method according to any one of claims 1-4, characterized in that, include: The module includes a model generation module, a perturbation module, a pricing module, and a dynamic update module. The model generation module is used to collect historical behavior data of users' logistics quotations and to build a user behavior model based on the historical behavior data. The disturbance module is used to identify user behavior patterns that affect the initial pricing strategy based on the user behavior model, and generate corresponding disturbance factors. The pricing module is used to adjust the initial pricing strategy based on the disturbance factor to obtain a pricing result that includes user behavior disturbances. The dynamic update module is used to dynamically update the user behavior model and disturbance factor based on the deviation between order fulfillment status and actual user behavior.

6. A logistics quotation management system as described in claim 5, characterized in that, The quotation module includes: a mapping unit, an adjustment unit, and a verification unit; The mapping unit is used to identify the behavior category labels in the disturbance factors and map them to a preset set of price adjustment parameters; The adjustment unit is used to adjust the quotation parameters during the quotation calculation stage; The verification unit is used to perform a global consistency check on the output of all computing nodes based on the disturbance factor before generating the quotation result, and to re-trigger the necessary quotation recalculation process according to the verification result.

Images