A multi-format service bill automatic processing method and system and a storage medium

Abstract

The present application relates to the technical field of logistics service, and more particularly to a multi-format service bill automatic processing method and system, and a storage medium, which first acquires multi-format logistics original communication bill data and each format business rule parameter; secondly, the two are matched to construct a matching matrix of the to-be-accounted bill item and the rule applicable range, and accordingly generate an account processing allocation strategy; finally, the strategy is coordinated and optimized based on an optimization objective function to determine a dynamic automatic processing strategy. The method solves the account distribution problem caused by complex billing logic and data heterogeneity in the multi-format cross-scene through the establishment of a precise mapping matrix of data and rules, and realizes the change from manual reconciliation to intelligent automation. At the same time, the dynamic optimization mechanism is introduced, which can adaptively adjust the strategy according to the multi-objective requirements such as error rate, timeliness and compliance, significantly improves the accuracy and efficiency of bill processing and the adaptability of the system to business changes, and effectively reduces the financial risk.

Description

Technical Field

[0001] This invention relates to the field of logistics service technology, and in particular to a method, system and storage medium for automatic processing of bills for multi-business services. Background Technology

[0002] With the rapid development of the logistics industry and the deepening of digital transformation, logistics companies generally involve multiple business forms, such as warehousing, trunk transportation, urban distribution, cross-border logistics, and cold chain services, forming a new service model of multi-business collaborative operation. Under this model, there are complex business overlaps and cost sharing relationships between various business forms, resulting in logistics service billing data with wide sources, diverse formats, and heterogeneous rules. Traditional manual reconciliation or automated processing methods based on fixed templates are no longer sufficient to meet the needs of efficient, accurate, and compliant financial settlement.

[0003] Currently, most enterprises still rely on manual experience to analyze and classify communication data from different systems. This is not only inefficient but also prone to errors, omissions, and duplicate billing due to misunderstandings of rules and inconsistent data definitions. Especially when dealing with massive amounts of high-frequency, real-time billing data, existing methods exhibit significant limitations in cross-industry identification, dynamic rule adaptation, and handling of abnormal entries. Furthermore, the billing logic varies greatly across different logistics sectors, such as billing by weight, volume, mileage, duration, or service level, further complicating automatic billing. While some research has attempted to introduce rule engines or machine learning models for billing classification, most are limited to single-industry scenarios, lacking unified semantic analysis capabilities for multi-source heterogeneous communication data and failing to fully consider the dynamic coupling between business rules and actual billing characteristics. Especially when handling disputed bills, shared service items, or situations requiring adjustments, the system lacks flexibility and struggles to achieve end-to-end intelligent billing decisions.

[0004] The above content is only used to help understand the technical solution of the present invention and does not represent an admission that the above content is prior art. Summary of the Invention

[0005] The main objective of this invention is to provide a method, system, and storage medium for automatic billing of multi-business services, aiming to solve the technical problems of existing logistics billing methods being unable to efficiently and accurately cope with the difficulties in automatic billing and the risk of errors and omissions caused by heterogeneous communication data, complex billing rules, and lack of dynamic coupling relationships under multi-business collaborative operation.

[0006] To achieve the above objectives, the present invention provides a method for automatically processing bills for multi-business services, the method comprising: Obtain raw communication billing data for multi-business logistics and business rule parameters for each business; The original communication billing data of the multi-format logistics and the business rule parameters of each format are matched to determine the matching matrix of the billing items to be split and the applicable scope of each format rule, and the splitting processing allocation strategy of the multi-format logistics service bill is determined according to the matching matrix. Based on the optimization objective function of the revenue sharing allocation strategy, the revenue sharing allocation strategy is coordinated and optimized to determine the dynamic automatic processing strategy.

[0007] Optionally, the matching matrix for determining the billing entries to be split and the applicable scope of rules for each business type includes: Based on the original communication billing data of the multi-business logistics, billing item features are extracted, and based on the business rule parameters of each business business, business rule adaptation features are extracted. Determine a first interaction relationship between the bill entry features and the business format rule adaptation features, and determine a coupling adaptation model between the bill entry and the business format rule based on the first interaction relationship; The coupling adaptation model is used to match the bill item features and the business rule adaptation features to determine the matching matrix between the bill items to be split and the applicable scope of each business rule.

[0008] Optionally, determining the revenue sharing strategy for multi-format logistics service bills based on the matching matrix includes: The characteristic type of the bill item is determined based on the original communication bill data of the multi-business logistics; Determine the characteristic quantification indicators of the bill item feature types and the business rule parameters of each business format; Based on historical settlement data, cross-business shared bills and deduction items are used to identify abnormal adjustment and supplementary entries. The abnormal adjustment supplementary entry is used as the basis for supplementary adjustment of the disputed bill period. The supplementary adjustment basis and the feature quantification index are matched by the matching matrix to determine the billing and allocation strategy for multi-format logistics service bills.

[0009] Optionally, determining the feature type of the bill item based on the original communication bill data of the multi-business logistics includes: Analyze the original communication billing data of the multi-business logistics to determine the second interaction relationship between the item attributes, amount scope, transaction scenario and business type attribution of the billing; Based on the second interaction relationship, determine the bill entry feature type.

[0010] Optionally, the characteristic quantification indicators for determining the feature type of the bill item and the business rule parameters of each business type include: Analyze the characteristic types of the bill items to determine quantitative indicators for bill attribution, quantitative indicators for bill time regularity, and quantitative indicators for bill amount fluctuation. Determine the third interaction relationship between the bill attribution quantitative indicator, the bill time regularity quantitative indicator, the bill amount fluctuation quantitative indicator, and the business rule parameters of each business format; The third interaction relationship is analyzed and quantified using cosine similarity to obtain characteristic quantitative indicators of business rules for each business format.

[0011] Optionally, the step of using the abnormal adjustment supplementary entry as the basis for supplementary adjustment of the disputed billing period, and determining the billing allocation strategy for multi-format logistics service bills by matching the supplementary adjustment basis and the feature quantification indicators through the matching matrix, includes: Obtain quantifiable indicators of bill attribution, bill time regularity, and bill amount fluctuation; Based on the bill attribution quantitative indicator, the bill time regularity quantitative indicator, and the bill amount fluctuation quantitative indicator, the accounting rule adaptation parameters, abnormal threshold parameters, and trigger parameters for using abnormal adjustment supplementary entries as supplementary adjustments for disputed bills are determined for the automatic processing of bills for multi-business logistics services. Based on the revenue sharing rule adaptation parameters, the anomaly threshold parameters, and the triggering parameters, a revenue sharing processing and allocation strategy for multi-business logistics service bills is determined.

[0012] Optionally, the optimization objective function based on the revenue sharing allocation strategy coordinates and optimizes the revenue sharing allocation strategy to determine a dynamic automatic processing strategy, including: The Pareto front resolution is dynamically adjusted based on the adaptive reference point and the scale of the logistics industry using the third-generation non-dominated sorting genetic algorithm, and the weight of different accounts is determined based on the time discounting effect of the third-generation non-dominated sorting genetic algorithm. Determine the conflict mechanism among multiple objectives in the objective function; Based on the adjusted Pareto front resolution and the different account weights, the objective function is dynamically optimized by balancing the conflict mechanisms between the objectives and determining the dynamic automatic processing strategy for the account allocation strategy.

[0013] Optionally, the objective function may include one or more of the following: the objective function for minimizing the error rate, the objective function for minimizing bill processing time, and the objective function for maximizing settlement compliance.

[0014] Furthermore, to achieve the above objectives, the present invention also provides an automatic billing system for multi-business services, the automatic billing system for multi-business services comprising: The data acquisition module is used to acquire original communication billing data for multi-business logistics and business rule parameters for each business. The matrix matching module is used to match the original communication billing data of the multi-business logistics and the business rule parameters of each business, determine the matching matrix of the billing items to be split and the applicable scope of each business rule, and determine the splitting processing and allocation strategy of the multi-business logistics service bill based on the matching matrix. The dynamic optimization module is used to coordinate and optimize the revenue sharing allocation strategy based on the optimization objective function of the revenue sharing allocation strategy, and determine the dynamic automatic processing strategy.

[0015] In addition, to achieve the above objectives, the present invention also provides a storage medium storing a multi-business service bill automatic processing program, which, when executed by a processor, implements the steps of the multi-business service bill automatic processing method as described in any of the above claims.

[0016] This invention provides an automated billing method for multi-business service formats. The method directly acquires and parses raw communication data from multiple business formats, abandoning the traditional reconciliation model that relies on manual experience or fixed templates. Utilizing matching matrix technology, it can automatically and accurately map massive, heterogeneous billing entries to complex business rules for each business format, achieving full automation from data access to billing strategy generation, significantly shortening the settlement cycle and reducing labor costs. Addressing the complex billing logic and inconsistent data definitions caused by the overlap of multiple business formats such as warehousing, transportation, and delivery in the logistics industry, this method constructs a deep matching mechanism between billing data and business rules. By determining the matching matrix between the billing entries to be billed and the applicable scope of the rules, it can intelligently identify cross-business shared services, special billing scenarios, and ambiguous entries, effectively avoiding errors caused by misunderstandings of rules. The method significantly improves the accuracy of financial data by addressing errors, omissions, and duplicate billing caused by differences in data formats. The introduction of a coordinated optimization mechanism based on an objective function allows the system to move beyond static rule execution and dynamically adjust billing strategies according to actual business needs. This dynamic adaptability ensures that the system can still output the optimal billing allocation scheme when facing fluctuations in business volume, rule changes, or disputed bills, enhancing the flexibility and robustness of financial management. This method decouples communication data parsing from business rule parameters, and through standardized matching matrices and optimization models, the system can easily adapt to new logistics business models or changing billing rules without refactoring core code. Furthermore, the dynamic automatic processing of strategy generation exhibits intelligent decision-making characteristics, providing solid technical support for logistics companies to transform towards intelligent and refined financial management. Attached Figure Description

[0017] Figure 1This is a flowchart illustrating an embodiment of the multi-business service bill automatic processing method of the present invention; Figure 2 This is a structural block diagram of an embodiment of the multi-business service bill automatic processing system of the present invention.

[0018] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0019] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0020] Reference Figure 1 , Figure 1 This is a flowchart illustrating an embodiment of the automatic billing method for multi-business services of the present invention, which presents an embodiment of the automatic billing method for multi-business services of the present invention.

[0021] In one embodiment, the method for automatically processing multi-business service bills includes: Step S100: Obtain the original communication billing data for multi-business logistics and the business rule parameters for each business business.

[0022] The multi-format logistics raw communication billing data can be unprocessed raw billing communication records from different logistics formats such as warehousing, trunk transportation, urban distribution, cross-border logistics, and cold chain services. It can serve as the original input data source for billing processing, carrying actual logistics service behaviors and billing information. In this embodiment, the multi-format logistics raw communication billing data can be directly collected from various business systems through API interfaces, file transfers, message queues, etc. For example, the multi-format logistics raw communication billing data can include, but is not limited to, one or more of the following: warehousing system billing data, trunk transportation system billing data, and urban distribution system billing data. The business rule parameters for each format can be a structured set of rules describing the billing logic, applicable conditions, and allocation ratios of different logistics formats. These can be used to provide a rule basis for billing item matching and billing strategy generation. Furthermore, the business rule parameters for each format can be provided by the business configuration center or rule management system, expressing the billing model in a parameterized form. In an exemplary embodiment, the business rule parameters for each format can include, but are not limited to, weight-based billing rule parameters, volume-based billing rule parameters, and service level-based billing rule parameters.

[0023] Obtaining raw communication billing data and business rule parameters for logistics across multiple business sectors can be achieved by synchronously pulling or listening to push raw billing data from multiple heterogeneous data sources, and loading currently effective business rule parameters from the rule base. Furthermore, this operation can be implemented by batch extracting historical billing data using ETL tools and loading static rule files, or by subscribing to billing streams and dynamically obtaining rule parameters through real-time message middleware. This enables unified access to data and rules, providing a foundation for subsequent automated processing.

[0024] Step S200: Match the original communication bill data of multi-business logistics with the business rule parameters of each business business to determine the matching matrix of the bill items to be split and the applicable scope of each business rule, and determine the split processing and allocation strategy of multi-business logistics service bills based on the matching matrix.

[0025] The billing entries to be allocated can be independent billing items extracted from the original communication billing data of multi-business logistics, requiring attribution determination and cost allocation. These can serve as the basic units for constructing the matching matrix and participate in the mapping process with business rules. In a specific embodiment, the billing entries to be allocated can include, but are not limited to, single-business-specific billing entries, cross-business-shared billing entries, and billing entries with fuzzy attribution. The scope of application for each business rule can be a set of boundary conditions defining the conditions under which the business rule parameters for each business can be applied. This can be used to limit the effective space for rule matching and avoid rule misuse. For example, the scope of application for each business rule can include, but is not limited to, time range, service area range, and customer type range.

[0026] A matching matrix can be a two-dimensional structured mapping table representing the matching relationship between bill entries to be allocated and the applicable scope of rules for each business type. It can be used to achieve accurate association between heterogeneous bills and complex rules, supporting cross-business type identification and fuzzy entry discrimination. In this embodiment, the matching matrix can calculate the fit between bill entry features and rule application conditions and quantify them into matrix elements through semantic parsing and rule condition comparison. Furthermore, the matching matrix works in conjunction with the bill entries to be allocated and the applicable scope of rules for each business type, with the former providing row indexes and the latter providing column indexes; the matching matrix also works in conjunction with the bill allocation strategy to provide an initial allocation basis. In an exemplary embodiment, the matching matrix may include, but is not limited to, a complete matching matrix, a partial matching matrix, and a fuzzy matching matrix. The bill allocation strategy can be a scheme for allocating bill expenses to relevant business types based on the initial determination of the matching matrix, which can be used as input for coordination optimization, reflecting the initial bill allocation intention after rule matching. For example, the bill allocation strategy may include, but is not limited to, a proportional allocation strategy, a primary responsibility allocation strategy, and a mixed allocation strategy.

[0027] Matching original communication billing data for multi-business logistics with business rule parameters for each business sector can involve feature extraction of billing entries and logical comparison or semantic similarity calculation with the applicable conditions of each business sector's rules. Further, this operation can be achieved through rule-engine-based conditional matching or vector embedding-based semantic similarity matching, thereby establishing a preliminary association between bills and rules and providing a basis for constructing the matching matrix. Determining the matching matrix for billing entries to be allocated and the applicable scope of each business sector's rules can involve pairwise evaluation of each billing entry with the applicable scope of all business sector rules, generating matrix elements representing the degree of matching. Further, this operation can be achieved by using Boolean matching to generate a 0-1 matrix or using membership functions to generate a continuous value matching matrix, thereby structurally expressing the complex mapping relationship between bills and rules and supporting cross-business and fuzzy scenario recognition. Based on the matching matrix, determining the allocation strategy for multi-business logistics service bills can be done by initially allocating billing fees to the corresponding business sector based on the position and strength of non-zero elements in the matching matrix. Furthermore, this operation can be achieved by uniquely attributing based on the maximum matching strength and by weighted apportionment based on matching strength, thereby generating an initial revenue sharing scheme that reflects the rule matching results.

[0028] Step S300: Based on the optimization objective function of the revenue sharing allocation strategy, coordinate and optimize the revenue sharing allocation strategy to determine the dynamic automatic processing strategy.

[0029] The optimization objective function can be a mathematical expression used to measure the merits of the revenue sharing and allocation strategy, reflecting the business optimization objective. It can drive the coordination and optimization process, guiding the system to output the optimal strategy that meets business requirements. In this embodiment, the optimization objective function can be transformed from business objectives (such as cost minimization and compliance maximization) into a computable objective function form. Furthermore, the optimization objective function and the revenue sharing and allocation strategy work together, with the former as the variable and the latter as the objective, for optimization and solution. For example, the optimization objective function can include, but is not limited to, cost minimization objective functions, dispute minimization objective functions, and rule consistency maximization objective functions. The dynamic automatic processing strategy can be a billing and revenue sharing execution scheme that is finally determined after coordination and optimization and has adaptive capabilities. It can be used as the system output result, directly used for financial settlement execution, and supports robust decision-making under rule changes and abnormal scenarios. In a specific embodiment, the dynamic automatic processing strategy can include, but is not limited to, steady-state routine processing strategies, fluctuation adaptation processing strategies, and anomaly fault-tolerant processing strategies. Furthermore, the dynamic automatic processing strategy is generated based on the optimization results of the optimization objective function and the revenue sharing and allocation strategy.

[0030] Based on the objective function of the revenue sharing allocation strategy, the strategy is coordinated and optimized. This can be achieved by treating the revenue sharing allocation strategy as a variable and solving for the optimal solution under the constraints of the objective function. Furthermore, this operation can be implemented using a linear programming solver for global optimization and heuristic algorithms for approximate optimization, thereby shifting the revenue sharing strategy from rule-driven to goal-driven, improving business adaptability. Determining the dynamic automatic processing strategy can be achieved by solidifying the coordinated optimization results into an executable set of revenue sharing instructions. Further, this operation can be implemented by generating JSON-formatted revenue sharing instruction packages and writing them to a database revenue sharing task table, thus outputting an adaptive final revenue sharing scheme that supports real-time settlement.

[0031] Taking the settlement of mixed cross-border cold chain and domestic trunk transportation orders as an example, the automatic billing method for multi-business service in this embodiment can be as follows: A logistics company accepts an order that is transported from an overseas warehouse to China via air freight and then delivered domestically via trunk line. The original communication bill includes three expense items: international segment temperature control service, customs clearance service, and domestic trunk transportation. The system obtains the bill and the rule parameters of the three business types: cross-border logistics, cold chain, and trunk transportation. Through the matching matrix, it is identified that the temperature control service meets both cold chain and cross-border rules, the customs clearance service only matches cross-border rules, and the trunk transportation only matches domestic rules. The initial revenue sharing strategy allocates the temperature control cost at 50-50. However, the optimization objective function is set to minimize disputes. The system finds that the customer's contract stipulates that the cross-border segment temperature control is borne by the shipper, so the allocation ratio is adjusted to 100-0. Finally, the dynamic automatic processing strategy outputs the revenue sharing result where the cross-border segment bears all temperature control costs, the customs clearance cost goes to the cross-border segment, and the trunk transportation cost goes to the domestic transportation segment.

[0032] In one embodiment, determining the matching matrix between the billing entries to be split and the applicable scope of rules for each business type includes: Based on the original communication billing data of logistics in multiple business formats, the features of billing items are extracted, and based on the business rule parameters of each business format, the business format rule adaptation features are extracted. Determine the first interaction relationship between bill item features and business format rule adaptation features, and determine the coupling adaptation model of bill items and business format rules based on the first interaction relationship; A coupled adaptation model is used to match the characteristics of bill items and the adaptation characteristics of business rules, and to determine the matching matrix between the bill items to be split and the applicable scope of each business rule.

[0033] The billing item features can be structured information vectors extracted from the original communication billing data of multi-business logistics, used to characterize the semantics and billing attributes of billing items. These vectors can serve as input elements in the matching process, supporting semantic alignment and logical comparison with rule conditions. In an exemplary embodiment, billing item features can extract key dimensions such as service type, billing unit, timestamp, and location from the original billing text or structured fields through natural language processing, field mapping, or rule guidance. Furthermore, billing item features can include, but are not limited to, one or more of service type features, billing dimension features, and spatiotemporal context features. Business format rule adaptation features can be a set of key attributes abstracted from the business rule parameters of various business formats, used to describe the applicable conditions and billing logic of the rules. These can be used as matching inputs on the rule side, forming a symmetrical comparison space with the billing item features. For example, business format rule adaptation features can extract computable elements such as applicable service scope, billing formula variables, and constraint thresholds by standardizing and parsing the business rule parameters. In a specific embodiment, business format rule adaptation features can include, but are not limited to, applicable condition features, billing formula features, and apportionment constraint features.

[0034] Based on raw communication billing data from various logistics sectors, feature extraction of billing items can be achieved by parsing and structuring the raw billing data to identify and extract semantic or numerical features related to billing. Furthermore, this operation can be implemented by using predefined field mapping templates to extract structured fields and employing named entity recognition models to extract service and metering information from free text, thereby transforming unstructured or heterogeneous bills into a unified feature representation and providing standardized input for subsequent matching. Based on business rule parameters for each sector, sector rule adaptation features can be extracted by semantically deconstructing the business rule parameters to extract rule condition elements that can be used for matching calculations. Further, this operation can be achieved by parsing rule syntax trees to extract constraint variables and using knowledge graphs to transform rule clauses into attribute triples, thereby transforming rules from logical descriptions into computable features and achieving standardized expression on the rule side.

[0035] The first interaction relationship can be a semantic, logical, or numerical association pattern between bill item features and business rule adaptation features. It can be used to provide a basis for constructing a coupled adaptation model, reflecting the dynamic coupling mechanism between features. In this embodiment, the first interaction relationship can quantify the degree of fit between the two through similarity calculation, logical implication judgment, or dimensional mapping analysis. For example, the first interaction relationship can include, but is not limited to, semantic similarity relationships, logical inclusion relationships, and numerical mapping relationships. Determining the first interaction relationship between bill item features and business rule adaptation features can be achieved by calculating the association strength or compatibility of the two types of features in the semantic or logical space. Furthermore, this operation can reveal the potential matching logic between bills and rules by calculating vector cosine similarity as semantic association and using a logical rule engine to determine whether conditions are met, thus going beyond literal matching.

[0036] The coupling adaptation model can be a computational model built based on the first interaction relationship to evaluate the matching degree between bill items and business rules. It can be used to achieve deep and dynamic matching of bills and rules, supporting intelligent discrimination in fuzzy, cross-domain, and partially missing scenarios. In a specific embodiment, the coupling adaptation model can utilize machine learning or rule fusion methods to model the interaction relationship between features as a generalizable matching function. Further, the coupling adaptation model can receive bill item features and business rule adaptation features as input and output a matching score for generating a matching matrix. For example, the coupling adaptation model can employ a neural network-based coupling adaptation model, a graph inference-based coupling adaptation model, or a rule embedding-based coupling adaptation model. Determining the coupling adaptation model of bill items and business rules based on the first interaction relationship can be achieved by using the interaction relationship as a training signal or construction basis to form a reusable matching model. Further, this operation can be implemented by using supervised learning to train a classification model to predict matching probabilities and constructing a feature interaction network based on an attention mechanism, thereby establishing a dynamic matching mechanism with generalization capabilities to adapt to changes in rules and data. The coupling adaptation model is used to establish the mapping relationship between bill entries and rules, and its expression is: In the formula, This is a matching matrix between the bill entries to be split and the applicable rules for each business type. The output indicates which bill entries are subject to which rules. Features of bill entries; Features adapted to business model rules; This is the first interaction function, representing the deep-level correlation logic between bill features and rule features; This is a quantitative indicator based on the cosine similarity calculated above; as well as These are weighting coefficients used to adjust the proportion of interaction relationships and similarity in the model; This represents biases that the model failed to explain. The expression uses the model to comprehensively process feature extraction and relationship matching, generating the final matching matrix as the basis for subsequent revenue sharing strategies. Using a coupled adaptation model to match bill item features and business rule adaptation features can be achieved by inputting the two types of extracted features into the coupled adaptation model and outputting a matching score or confidence level. Furthermore, this operation can be implemented through batch inference to generate a matching score matrix or online real-time model invocation for single-item matching, thus achieving high-precision, flexible bill-rule matching and supporting fuzzy and cross-domain scenarios. The matching matrix that determines the applicable scope of the bill items to be shared and each business rule can be a two-dimensional matrix structure that summarizes the matching scores of all bill items and each business rule. Furthermore, this operation can be achieved by filling matrix elements with probability values ​​and using Boolean values ​​plus weighting coefficients to represent matching strength, thus forming a structured matching result as the basis for generating revenue sharing strategies.

[0037] Taking the automatic classification of fuzzy descriptions of bills in cold chain transportation as an example, the automatic processing method for multi-business service bills in this embodiment can be as follows: A bill item is described as a temperature-controlled delivery service throughout the entire process, without clearly distinguishing between international and domestic segments; the system extracts the service type as temperature-controlled, the billing unit as mileage, and the spatiotemporal context as a cross-border route from the item; the system extracts the adaptation features of the cold chain business from the business rules as a temperature range of 2-8 degrees Celsius, a billing dimension including mileage and duration, and an applicable area including cross-border; through the coupling adaptation model calculation, it is found that the item has a high semantic similarity to the cross-border cold chain rules and meets the logical conditions, but there is a regional mismatch with the pure domestic cold chain rules; the first interaction relationship is reflected in partial logical inclusion and high semantic overlap; the final matching matrix assigns high confidence to the cross-border cold chain rule column and low weight to the domestic cold chain column, supporting reasonable allocation in the subsequent optimization stage.

[0038] In one embodiment, the revenue sharing strategy for multi-format logistics service bills is determined based on the matching matrix, including: The characteristic types of bill items are determined based on the original communication bill data of multi-business logistics.

[0039] In this context, the bill entry feature type can be a category label abstracted and categorized from the original communication billing data of multi-business logistics based on service attributes, billing dimensions, or business affiliation. This label can provide structured input for subsequent rule matching and quantitative analysis, improving matching accuracy. In this embodiment, the bill entry feature type can be generated through semantic parsing and pattern recognition of the field content, service description, and billing unit of the bill entry. For example, the bill entry feature type can include, but is not limited to, single service feature type, composite service feature type, and shared service feature type. Determining the bill entry feature type based on the original communication billing data of multi-business logistics can involve parsing the text description, service code, billing unit, and other fields of the original bill entry and classifying them into a predefined feature type system. Furthermore, this operation can be achieved through keyword rule template matching of feature types and semantic classification based on a pre-trained language model, thereby realizing the structured representation of unstructured billing data and laying the foundation for subsequent quantification and matching.

[0040] Determine the characteristic quantification indicators of the bill item feature types and business rule parameters for each business format.

[0041] The feature quantification index can be a calculable numerical metric that transforms the qualitative conditions in the bill item feature types and business rule parameters of each business type into a computable numerical standard. This standard can be used to support automated comparison in the matching matrix based on numerical similarity or logical thresholds. In an exemplary embodiment, the feature quantification index can convert non-numerical features such as text and categories into vectors or scalars through a preset mapping function or embedding model. For example, the feature quantification index can include, but is not limited to, Boolean indicators, continuous indicators, and discrete-level indicators. Determining the feature quantification index for the bill item feature types and business rule parameters of each business type can be achieved by mapping the bill feature types and the condition items in the rule parameters to numerical vectors or indicator sets of a unified dimension. Furthermore, this operation can be implemented using one-hot encoding to generate discrete indicators and using an embedding layer to generate dense vector indicators, thereby establishing a computable alignment space between bills and rules, supporting fine-grained matching.

[0042] Based on historical settlement data, cross-business shared bills and deduction items are used to identify abnormal adjustment and supplementary entries.

[0043] Historical settlement data can be a collection of invoice records from a company's past financial settlements, including accounting results, adjustment records, and dispute resolution information. This data can serve as the foundation for anomaly pattern identification and supplementary entry generation. In one specific embodiment, historical settlement data can be extracted from archived settlement details in a financial system or data warehouse. For example, historical settlement data may include, but is not limited to, regular settlement records, dispute mediation records, and manually entered records. Cross-industry shared invoices can be invoice entries in historical settlement data where costs are shared by multiple logistics businesses. This data can be used to identify service scenarios requiring joint allocation and guide the generation of anomaly adjustment supplementary entries. For example, cross-industry shared invoices may include, but are not limited to, shared invoices for warehousing and distribution, shared invoices for trunk lines and cold chain logistics, and shared invoices for cross-border and customs clearance.

[0044] Deduction items can be billing items that have been reduced or waived in historical settlements due to discounts, rebates, breach of contract compensation, etc. They can be used to reflect non-standard billing behavior and serve as important clues for abnormal adjustments. For example, deduction items can include, but are not limited to, contract discount deductions, service quality deductions, and customer rebate deductions. Abnormal adjustment supplementary entries can be supplementary adjustment rules or parameters identified based on historical settlement data to correct special billing situations that cannot be covered by regular rules. They can be used as a basis for dynamic correction of disputed billing periods, enhancing the contextual awareness of billing strategies. In an exemplary embodiment, abnormal adjustment supplementary entries can be used to analyze patterns of shared billing and deduction items across business formats to generate reusable adjustment factors. Furthermore, abnormal adjustment supplementary entries can be coordinated with disputed billing periods to limit their applicable time range; and coordinated with the matching matrix to participate in strategy generation as an additional matching dimension. For example, abnormal adjustment supplementary entries can include, but are not limited to, proportion correction supplementary entries, attribution transfer supplementary entries, and fee waiver supplementary entries. Based on cross-business shared bills and deduction items from historical settlement data, abnormal adjustment and supplementary entries can be identified. This can be achieved by mining frequently occurring shared allocation patterns or deduction logics in historical settlements and extracting them into reusable supplementary entry rules. Furthermore, this operation can discover high-frequency sharing patterns through association rule mining and identify typical deduction scenarios through cluster analysis, thereby making implicit business experience explicit and forming a dynamic supplement to regular rules.

[0045] The abnormal adjustment supplementary entries are used as the basis for supplementary adjustments to the disputed billing period. By matching the supplementary adjustment basis and characteristic quantitative indicators through the matching matrix, the accounting and allocation strategy for multi-format logistics service bills is determined.

[0046] The disputed billing period can be a time interval within the billing cycle where there is unclear attribution, billing conflict, or customer objection. It can be used to define the effective scope of abnormal adjustment supplementary entries, avoiding overgeneralization. For example, the disputed billing period can include, but is not limited to, customer complaint periods, system failure periods, and rule transition periods. The supplementary adjustment basis can be a structured adjustment rule or weight parameter derived from the abnormal adjustment supplementary entry and applicable to the current bill matching. It can be used to inject historical experience into the matching process, improving the ability to handle ambiguous or abnormal bills. For example, the supplementary adjustment basis can include, but is not limited to, adjustment coefficients, priority markers, and exclusion rules. Using abnormal adjustment supplementary entries as supplementary adjustment basis for disputed billing periods can be achieved by activating supplementary entries within the corresponding period based on the current bill's timestamp or dispute identifier and converting them into matching-usable adjustment parameters. Furthermore, this operation can be implemented by registering supplementary entries as rule plugins with time windows and constructing a disputed period-supplementary entry mapping index table, thereby enabling accurate reuse of historical experience in specific contexts and avoiding global rule pollution.

[0047] By using a matching matrix to supplement the adjustment criteria and feature quantification indicators, the dimensions of the original matching matrix can be expanded to include joint matching calculations of the supplementary adjustment criteria and feature quantification indicators. Furthermore, this operation can add a weight column for supplementary entries to the matching matrix for weighted summation, and construct multi-channel matching sub-matrices to handle rules and supplementary entry criteria separately. This allows for the integration of rule-driven and data-driven judgments, improving the accuracy of handling complex cross-services and disputed bills. Determining the billing allocation strategy for multi-format logistics services can be achieved by comprehensively matching the original matching results and the supplementary adjustment criteria to generate the final cost allocation scheme. Further, this operation can employ a weighted voting mechanism to integrate multi-source matching results and integrate supplementary adjustment constraints through constraint optimization. This allows for the output of a billing strategy that balances rule compliance and historical experience adaptability, enhancing robustness.

[0048] Taking the handling of disputed bills in urban delivery and warehousing joint services as an example, the automatic billing method for multi-business services in this embodiment can be as follows: A customer uses both warehousing and same-day delivery services during a promotional period, and a high loading and unloading fee appears on the bill. The system initially matches this as meaning that the entire fee should be borne by the warehousing department. However, historical settlement data shows that in similar scenarios, this fee is often shared between delivery and warehousing at a ratio of 6:4, and there are customer deduction records. The system generates an abnormal adjustment supplementary entry based on historical cross-business shared bills and deduction entries, indicating that this type of loading and unloading fee should be shared proportionally during the promotional period. The current bill time falls within the disputed bill time period, and this supplementary entry is activated as a supplementary adjustment basis. When calculating the matching matrix, both the original rule matching and the supplementary entry basis are considered simultaneously, and the final output is a 6:4 billing allocation strategy.

[0049] In one embodiment, the characteristic type of the bill item is determined based on the original communication billing data of multi-business logistics, including: Analyze the original communication billing data of multi-business logistics to determine the second interaction relationship between the item attributes, amount scope, transaction scenario and business type attribution of the billing; Based on the second interaction relationship, determine the feature type of the bill entry.

[0050] The bill entry attributes can be fields describing service content, billing items, or operation types, such as service name, operation type, and billing item code. These attributes reflect the basic business meaning of the bill entry and are one of the fundamental dimensions for feature recognition. In this embodiment, the bill entry attributes may include, but are not limited to, service identifier attributes, operation action attributes, and billing item attributes. The amount scope can be the billing benchmark or calculation logic description corresponding to the amount value in the bill entry, such as tax-inclusive price, excluding surcharges, or volume-based conversion. This can be used to define the business meaning boundary of the amount and avoid billing errors due to misunderstandings. For example, the amount scope can adopt the basic service amount scope, the surcharge amount scope, or the net amount after discount. The transaction scenario can be the specific business context in which the bill occurs, including customer type, promotional activities, service hours, geographical region, and other situational factors. This can provide external constraints on the bill semantics and affect the actual applicability of the billing rules. In an exemplary embodiment, the transaction scenario may include customer level scenarios, holiday operation scenarios, and emergency dispatch scenarios. Business type attribution refers to the business category to which a billing item belongs within the multi-business system of logistics, such as warehousing, long-haul transportation, and urban delivery. It is used to determine which business rules should apply to the bill and is the core basis for billing attribution. Furthermore, business type attribution can include attribution to front-end pickup, mid-stream transportation, and last-mile delivery.

[0051] The second interaction relationship can be a non-linear, coupled semantic association pattern, existing among the four elements of the bill: item attributes, amount scope, transaction scenario, and business type affiliation. In a specific embodiment, the second interaction relationship can be obtained by joint statistical analysis or graph neural network modeling of original communication bill data from multiple business types, to uncover implicit dependencies between dimensions. The second interaction relationship can be used to reveal the deep business semantics of the bill, supporting high-precision feature type determination. For example, the second interaction relationship can include conditional dependency interaction, semantic ambiguity resolution interaction, and cross-dimensional collaborative interpretation interaction. In this embodiment, the second interaction relationship integrates the joint semantics of item attributes, amount scope, transaction scenario, and business type affiliation as a direct basis for generating bill item feature types. Analyzing original communication bill data from multiple business types to determine the second interaction relationship between the bill's item attributes, amount scope, transaction scenario, and business type affiliation can be achieved by jointly modeling the four dimensional fields in the bill data to identify their non-independent, non-linear semantic coupling patterns. Furthermore, this operation can be achieved by constructing a four-dimensional cross-statistic table to identify high-frequency co-occurrence patterns and using graph neural networks to model semantic dependencies between dimensions, thereby breaking through the limitations of traditional isolated dimension processing and achieving a deeper understanding of the business intent of the bill.

[0052] Based on the second interaction relationship, the feature type of the bill item is determined. This can be achieved by mapping the identified interaction patterns to structured feature category labels or embedding vectors. In one specific embodiment, this operation can be implemented by converting the interaction patterns into predefined feature types based on rule templates and automatically discovering new feature types from the interaction space through clustering algorithms. This results in the generation of context-aware bill feature representations, supporting accurate bill allocation. The bill item feature type can be a composite bill semantic category generated based on the second interaction relationship and incorporating multi-dimensional contextual information. Furthermore, the bill item feature type can include single-dimensional dominant feature types, multi-dimensional coupled feature types, context-sensitive feature types, etc. In this embodiment, the bill item feature type is generated based on the second interaction relationship and serves as the input unit for subsequent matching and quantization processes.

[0053] Taking the confusion between fuel surcharges and temperature control service fees in cold chain transportation as an example, the automatic billing method for multi-business service in this embodiment can be as follows: A bill item shows an amount of 800 yuan, the item attribute is surcharge, the amount includes fuel surcharge, the transaction scenario is the summer high-temperature period, and the business type is cold chain transportation; the system analysis finds that under the cold chain business type, if the surcharge during the high-temperature period includes fuel but there is no actual trunk mileage record, it is likely to be a temperature control service fee; this judgment relies on the second interaction relationship between the item attribute, the amount, the transaction scenario, and the business type; the generated bill item feature type is high-temperature temperature control surcharge service, rather than ordinary fuel surcharge, thereby ensuring that the fee is correctly attributed to cold chain service rather than trunk transportation.

[0054] In one embodiment, the characteristic quantification indicators for determining the bill item feature type and business rule parameters for each business type include: Analyze the characteristic types of bill items to determine quantitative indicators for bill attribution, quantitative indicators for bill time regularity, and quantitative indicators for bill amount fluctuation. Determine the third interaction relationship between quantitative indicators of bill attribution, quantitative indicators of bill time regularity, quantitative indicators of bill amount fluctuation, and business rule parameters of each business format; By analyzing and quantifying the third interaction relationship using cosine similarity, characteristic quantitative indicators of business rules for each industry format are obtained.

[0055] Among them, the bill attribution quantification index can be a numerical measure characterizing the degree of attribution tendency of bill items across multiple business formats, reflecting the strength of its fit with the service attributes of a specific business format. It can be used to measure the probability that a bill should belong to a certain business format, supporting cross-business format identification. In an exemplary embodiment, the bill attribution quantification index can be weighted and scored or classified based on fields such as service description, operating entity, and location in the bill item feature type. Further, the bill attribution quantification index can include, but is not limited to, service entity attribution index, operating location attribution index, and customer contract attribution index. The bill time regularity quantification index can be a statistical measure depicting whether the bill occurrence time conforms to the typical business cycle or service window of a certain business format. It can be used to assist in identifying bills during abnormal periods or verifying the authenticity of services, improving the context awareness capability of rule matching. For example, the bill time regularity quantification index can calculate a regularity score by analyzing the time distribution pattern of the bill timestamp and historical similar bills (such as daily average, weekly peak, and holiday effect). In a specific embodiment, the bill time regularity quantification index can include, but is not limited to, daily cycle regularity index, weekly cycle regularity index, and seasonal regularity index.

[0056] Bill amount fluctuation quantification indicators can be standardized metrics that measure the degree of deviation of the bill amount from the historical average or expected value of the service type. They can be used to identify abnormal billing, discount deductions, or data entry errors, serving as a correction signal for rule adaptation. In this embodiment, the bill amount fluctuation quantification indicator can be calculated based on sliding window statistics or the Z-score method to determine the relative fluctuation level of the current amount compared to its historical distribution. Furthermore, bill amount fluctuation quantification indicators can include, but are not limited to, abnormal unit price fluctuation indicators, abnormal quantity fluctuation indicators, and combined service premium fluctuation indicators. Analyzing bill item feature types to determine bill attribution quantification indicators, bill time regularity quantification indicators, and bill amount fluctuation quantification indicators can be achieved by extracting key dimensions from the identified bill item feature types and calculating their quantitative values ​​in terms of attribution tendency, time regularity, and amount stability. Further, this operation can be implemented by generating three indicators using a predefined scoring card model and jointly outputting the indicators through a time-series model and anomaly detection algorithm, thereby transforming abstract features into computable multidimensional indicators and providing a structured foundation for high-order rule matching.

[0057] The third interaction relationship can be a high-dimensional coupled structure formed between three types of quantitative indicators—billing attribution, time regularity, and amount fluctuation—and business rule parameters for various business formats. This structure can transcend single-field matching and achieve deep alignment between billing behavior patterns and rule semantic logic. In a specific embodiment, the third interaction relationship can combine the three types of billing indicators with the conditional terms in the rule parameters using a Cartesian product to construct a multi-dimensional feature interaction space. For example, the third interaction relationship can include, but is not limited to, linear combination relationships, conditional dependencies, and non-linear coupling relationships.

[0058] Determining the third interaction relationship between bill attribution quantification indicators, bill time regularity quantification indicators, bill amount fluctuation quantification indicators, and business rule parameters for each business format can be achieved by combining the three bill quantification indicators with the applicable conditions in the rule parameters of each business format to form a high-dimensional interaction feature. Further, this operation can be implemented by constructing a tensor-form interaction matrix and generating combined feature vectors using a feature cross-layer, thereby establishing a deep mapping between bill behavior patterns and rule semantic logic, overcoming the limitations of field-level matching. Cosine similarity can be a standardized measure of the directional consistency between two vectors, with a value ranging from -1 to 1. It can be used to evaluate the similarity between bill behavior patterns and rule semantic patterns while ignoring absolute magnitude differences, improving the matching robustness under heterogeneous data. In this embodiment, cosine similarity can represent the third interaction relationship as a vector form and calculate the cosine value of the angle between it and the rule vectors of each business format. For example, cosine similarity can include, but is not limited to, original feature cosine similarity, weighted feature cosine similarity, and dimensionality-reduced cosine similarity.

[0059] The third interaction relationship is analyzed and quantified using cosine similarity to obtain the characteristic quantification index of business rules for each business type. This can be achieved by vectorizing the third interaction relationship and calculating the cosine similarity with the standard vector of each business type rule, outputting a matching score. Furthermore, this operation can be implemented by directly calculating the cosine value of the original high-dimensional vector, first performing PCA dimensionality reduction, and then calculating the cosine similarity. This enables robust semantic matching of heterogeneous and unaligned data, enhancing the system's adaptability in scenarios with format differences and inconsistent units. The expression used to quantify the degree of matching between bill item features and business rules is as follows: In the formula, The feature vector representing the bill item consists of quantitative indicators such as bill attribution, time regularity, and amount fluctuation, and is used to characterize the specific data features of the bill to be allocated. The feature vector representing the business rule parameters of each business type is composed of the rule parameters of the corresponding business type and serves as a benchmark for matching. , For vectors and The specific value in the i-th feature dimension; The cosine similarity value, ranging from [-1, 1], is used to calculate the cosine of the angle between two vectors, determining whether the bill features and business rules are consistent. The closer the value is to 1, the more consistent the feature directions are, and the higher the matching degree, thus transforming the qualitative matching relationship into a quantitative feature quantification indicator. The feature quantification indicator for each business rule can be a standardized value obtained after cosine similarity measurement, representing the matching strength between each business rule and the current bill, which can be used to support the accurate generation of subsequent billing strategies. In an exemplary embodiment, the feature quantification indicator for each business rule may include, but is not limited to, single rule matching score, rule cluster comprehensive score, and cross-business comparison score.

[0060] Taking the cross-business identification of cold chain nighttime delivery bills as an example, the automatic processing method for multi-business service bills in this embodiment can be as follows: If a bill shows a high temperature-controlled transportation fee incurred at 2:00 AM, the system first identifies its dual attributes of cold chain and time-sensitive delivery based on feature type. The calculated bill attribution quantitative index is biased towards cold chain, the bill time regularity quantitative index shows a significant deviation from the regular delivery period but conforms to the cold chain nighttime operation mode, and the bill amount fluctuation quantitative index shows a slight increase above the average but is a reasonable premium. The system constructs a third interaction relationship vector between these three indicators and the two types of rules: cold chain and urban delivery. Through cosine similarity calculation, it is found that the similarity with the cold chain rule vector is 0.89, which is much higher than the delivery rule's 0.42. Finally, the generated business rule feature quantitative indexes of each business type guide the matching matrix to completely attribute the bill to the cold chain business type, avoiding misjudgment as ordinary delivery due to time anomalies.

[0061] In one embodiment, the abnormal adjustment supplementary entry is used as the basis for supplementary adjustment of the disputed billing period. The supplementary adjustment basis and characteristic quantitative indicators are matched using a matching matrix to determine the billing allocation strategy for multi-format logistics services, including: Obtain quantifiable indicators of bill attribution, bill time regularity, and bill amount fluctuation; Based on quantitative indicators of bill attribution, quantitative indicators of bill time regularity, and quantitative indicators of bill amount fluctuation, we determine the matching parameters of the billing rules for automatic processing of bills for multi-business logistics services, the abnormal threshold parameters, and the trigger parameters for using abnormal adjustment supplementary entries as supplementary adjustments for disputed bills. Based on the revenue sharing rule adaptation parameters, abnormal threshold parameters, and trigger parameters, determine the revenue sharing processing and allocation strategy for multi-business logistics service bills.

[0062] Among them, the bill attribution quantification index can be a numerical measure of the strength of the association between a bill item and a specific logistics business format or service entity. It can be used to provide a basis for adapting billing rules and support bias judgment in fuzzy attribution scenarios. In this embodiment, the bill attribution quantification index can be calculated based on information such as the characteristic type of the bill item, service path, and operating entity, through statistical models or graph relationships to calculate the attribution probability or weight. Further, the bill attribution quantification index can include, but is not limited to, service node attribution index, operating entity attribution index, and contract liability attribution index. The bill time regularity quantification index can be a numerical indicator that characterizes whether the bill occurrence time conforms to historical periodicity, seasonality, or business rhythm patterns. It can be used to identify bills during abnormal periods (such as system delayed supplementation, temporary expedited services) and assist in dispute determination. For example, the bill time regularity quantification index can be used to assess the degree of deviation between the current bill time point and the historical distribution through time series analysis, Fourier transform, or sliding window statistics. In an exemplary embodiment, the bill time regularity quantification index can include, but is not limited to, periodic regularity index, sudden deviation index, and holiday effect index.

[0063] Bill amount fluctuation quantification indicators can be standardized measures reflecting the deviation of the current bill amount from the average or expected value of similar historical bills. They can be used to detect abnormally high or low bills, triggering manual review or adjustment mechanisms. In one specific embodiment, bill amount fluctuation quantification indicators can calculate Z-scores, coefficients of variation, or quantile distances based on the historical amount distribution within a sliding window. Furthermore, bill amount fluctuation quantification indicators can include, but are not limited to, positive surge indicators, negative anomaly indicators, and stability maintenance indicators.

[0064] Obtaining quantifiable indicators for bill attribution, bill time regularity, and bill amount fluctuation can be achieved by extracting features from historical and current bill data and calculating the values ​​of these three types of indicators. Furthermore, this operation can be implemented by calculating indicators based on statistical distributions or predicting indicators based on machine learning models, thereby building a structured perception capability of multi-dimensional behavioral characteristics of bills and providing input for dynamic parameter generation. Billing rule adaptation parameters can be adjustable parameters used to dynamically adjust the weight or priority of business rules for different business formats in the current bill processing. This can be used to achieve contextual adaptation of rule application and avoid one-size-fits-all matching. In this embodiment, the billing rule adaptation parameters can be derived from a comprehensive analysis of the three types of quantitative indicators: bill attribution, time, and amount, controlling the sensitivity of rule matching. For example, billing rule adaptation parameters may include, but are not limited to, rule weight coefficients, applicable priority numbers, and condition relaxation factors.

[0065] An anomaly threshold parameter can serve as a boundary value for determining whether a billing behavior has entered an abnormal state. It is used to trigger special processing procedures and can be used to accurately identify abnormal bills requiring intervention, reducing false alarms and missed alarms. In one exemplary embodiment, the anomaly threshold parameter can be dynamically set based on historical data distribution and can automatically drift with business seasons or scale. A trigger parameter can be an activation condition parameter that determines whether to enable anomaly adjustment supplementary entries as a basis for supplementary adjustments. It can be used to control the timing of the anomaly adjustment mechanism's intervention, ensuring that supplementary entries are introduced only when necessary. In a specific embodiment, the trigger parameter can be generated by a logical combination function when the bill attribution, time, or amount indicator exceeds the corresponding anomaly threshold. For example, the trigger parameter can include, but is not limited to, single indicator exceeding the limit trigger, multi-indicator joint trigger, and time-limited trigger.

[0066] Based on quantitative indicators of bill attribution, bill time regularity, and bill amount fluctuation, the system determines the adaptation parameters for the automatic processing of multi-format logistics service bills, anomaly threshold parameters, and trigger parameters for using anomaly adjustment supplementary entries as supplementary adjustments for disputed bills. This can be achieved by inputting the three types of quantitative indicators into a preset mapping function or decision model, outputting three types of control parameters. Furthermore, this operation can be implemented by configuring mapping tables to generate parameters through a rule engine or by generating parameters through lightweight neural network regression, thereby enabling the system to have context awareness and achieve automatic transformation from data features to processing strategies. Based on the adaptation parameters of the billing rules, anomaly threshold parameters, and trigger parameters, the system determines the billing processing allocation strategy for multi-format logistics service bills. This can be achieved by dynamically applying the above parameters to adjust rule weights, determine abnormal states, and decide whether to introduce anomaly adjustment supplementary entries during the matching matrix construction and strategy generation process. Furthermore, this operation can be achieved by weighting rule scores in the matching matrix calculation or by conditionally injecting supplementary entries during the strategy generation stage, thereby outputting an adaptive billing scheme that integrates real-time context and historical experience, improving accuracy and robustness.

[0067] Taking the abnormal handling of cold chain delivery bills during a major promotion as an example, the automatic processing method for multi-business service bills in this embodiment can be as follows: During a major promotion of an e-commerce platform, a cold chain delivery bill is generated, with the amount being 300% higher than usual, and the service time falling in the atypical delivery period in the early morning; the system calculates the bill amount fluctuation quantification index as 2.8 (exceeding the threshold of 2.0), and the bill time regularity quantification index shows a significant deviation from the daily distribution; at the same time, the bill involves multiple temperature zone services, and the bill attribution quantification index shows a multi-business equilibrium state; based on the three types of indicators, the system generates a lower trunk transportation rule weight and a higher abnormal threshold trigger signal, and activates the abnormal adjustment supplementary entry item for temporary price increases during the major promotion; the matching matrix integrates this supplementary entry item during calculation, and finally distributes the excess cost to the e-commerce customer and the cold chain service provider according to the temporary agreement ratio, avoiding the logistics party bearing all the costs according to the conventional rules.

[0068] In one embodiment, based on the optimization objective function of the revenue sharing allocation strategy, the revenue sharing allocation strategy is coordinated and optimized to determine a dynamic automatic processing strategy, including: The Pareto front resolution is dynamically adjusted based on the adaptive reference point and the scale of the logistics industry using the third-generation non-dominated sorting genetic algorithm, and the weight of different accounts is determined based on the time discounting effect of the third-generation non-dominated sorting genetic algorithm. Determine the conflict mechanism among multiple objectives in the objective function; Based on the adjusted Pareto front resolution and different account weights, the objective function is dynamically optimized by balancing the conflict mechanism between objectives and determining the dynamic automatic processing strategy for the account allocation strategy.

[0069] The third-generation non-dominated sorting genetic algorithm can be an evolutionary algorithm for solving high-dimensional multi-objective optimization problems. By guiding the population distribution on the Pareto front through reference points, it can serve as a computational engine for coordinated optimization, handling multiple conflicting business objectives in revenue sharing. In this embodiment, the third-generation non-dominated sorting genetic algorithm can be categorized by reference point generation method, including fixed uniform reference points, adaptive clustering reference points, and dynamic reference points based on historical solution sets. Adaptive reference points can be reference direction vectors dynamically adjusted according to the current population distribution in the third-generation non-dominated sorting genetic algorithm. They can be used to guide the search process to focus on insufficiently explored target regions, improving solution set diversity. For example, adaptive reference points can be reconstructed based on the density and empty regions of discovered non-dominated solutions in the target space. Furthermore, adaptive reference points can be categorized by adjustment strategy, including density-driven reference points, boundary expansion reference points, and target preference-guided reference points.

[0070] The scale of the logistics industry can be a comprehensive measure of the number of logistics industry types currently participating in revenue sharing and their business complexity. It can be used as an input parameter for adjusting the Pareto front resolution, reflecting the problem dimension and search difficulty. In an exemplary embodiment, the scale of the logistics industry can be classified according to combinatorial complexity, including single-industry, multi-industry coupling, and full-industry collaboration. The Pareto front resolution can be the minimum resolvable distance between adjacent solutions in the Pareto optimal solution set in the target space. It can be used to control the balance between the fineness of the optimization results and computational overhead. In a specific embodiment, the Pareto front resolution can be classified according to granularity, including coarse-grained resolution, medium-grained resolution, and fine-grained resolution.

[0071] The time discounting effect can be seen as an economic principle in finance that assigns different value weights to bills occurring at different points in time. It can be used to transform the proximity of payment periods into weighting factors in optimization objectives, reflecting the timeliness of cash flow. Furthermore, the time discounting effect can be categorized by discounting model, including exponential discounting, hyperbolic discounting, and stepped discounting. Different payment period weights can be numerical importance coefficients assigned to each bill's payment period based on the time discounting effect. This can be used to differentiate between near-term and far-term bills in multi-objective optimization, improving financial realism. In this embodiment, different payment period weights can be calculated using a discount function based on the time difference between the bill's occurrence date and the settlement benchmark date. The conflict mechanism between multiple objectives can be a structured expression describing the inverse relationship between sub-objectives (such as cost, compliance, and timeliness) in the optimization objective function. It can be used to explicitly model the trade-offs between objectives, providing a basis for conflict identification in dynamic optimization. In a specific embodiment, the conflict mechanism between multiple objectives can be categorized by conflict type, including resource competition conflicts, rule constraint conflicts, and priority opposition conflicts.

[0072] The adjusted Pareto front resolution can be a Pareto front solution set density control parameter dynamically modified by the scale of logistics business types. This parameter can be used to maintain a balance between solution set quality and computational efficiency under different business type complexities. Furthermore, the adjusted Pareto front resolution can be coordinated with the scale of logistics business types, with the former as the output and the latter as the input; and it can also be coordinated with the third-generation non-dominated sorting genetic algorithm as its internal parameter adjustment mechanism. The adaptive reference points based on the third-generation non-dominated sorting genetic algorithm and the dynamic adjustment of the Pareto front resolution based on the scale of logistics business types can be achieved by assessing the problem complexity based on the current scale of logistics business types and adjusting the reference point density accordingly, thereby changing the expected resolution of the Pareto front. Furthermore, this operation can be achieved by linearly scaling the number of reference points according to the number of business types, or dynamically setting the resolution threshold based on historical optimization performance feedback, thus ensuring that multi-objective optimization maintains the representativeness and convergence of the solution set even when the number of business types changes.

[0073] Determining the weighting of different accounts receivable based on the time discounting effect of the third-generation non-dominated sorting genetic algorithm can be achieved by mapping the account period of an account to a time discounting factor and embedding it into the weighting term of the objective function. Furthermore, this operation can be implemented by using a fixed discount rate to calculate weights and dynamically adjusting discounting parameters according to the company's cost of capital, thus ensuring that the optimization process prioritizes the accuracy and timeliness of recent accounts, aligning with the actual financial management practices of enterprises. Determining the conflict mechanism among multiple objectives of the objective function can be achieved by analyzing the gradient direction or sensitivity of each sub-objective and identifying their opposing relationships in the decision space. Furthermore, this operation can be implemented by identifying conflicts using the objective relevance matrix and inferring conflict intensity based on the Pareto solution set distribution, thus providing a priori conflict structure for subsequent trade-off optimization and avoiding ineffective searches.

[0074] Based on the adjusted Pareto front resolution and different account period weights, the objective function is dynamically optimized by balancing the conflict mechanisms between objectives. This can be achieved within the NSGA-III framework by integrating account period weights and conflict mechanism information to guide the population towards business-preferred regions. Furthermore, this operation can be implemented by introducing account period-weighted fitness into the selection operator and masking low-weight objective directions during the reference point allocation phase, thereby generating an account sharing strategy that satisfies both mathematical optimality and business semantic preferences. The expression for the collaborative optimization of the account sharing strategy based on the NSGA-III algorithm is as follows: Where the objective function Expanded to: In the formula, For the allocation strategy of revenue sharing, decision variables represent specific revenue sharing schemes such as cost sharing ratios and routing selection; F The Pareto optimal front solution set contains multiple conflicting objectives; The objective function is to minimize the error rate; it is a function that measures the accuracy of accounting. The lower the value, the better. The objective function is to minimize the time required for bill processing; it is a function that measures processing efficiency, and the lower the value, the better. The objective function for settlement compliance is to measure the degree of compliance. Here, we take the inverse as the minimization problem, where a lower value indicates higher compliance. The weighting of different payment periods is determined based on the time discounting effect, with more recent payments having a higher weight and longer payments having a higher weight; T represents the total number of payment periods considered. To constrain the strategy space and satisfy business logic limitations, such as the total amount must equal the total bill, a multi-objective optimization function is constructed. The NSGA-III algorithm is used to find the optimal balance between conflicting objectives (such as increased time consumption for compliance or increased error rate for speed), thereby determining the dynamic automatic processing strategy. Furthermore, the dynamic automatic processing strategy for determining the bill allocation strategy can be to select the final solution that conforms to current business preferences from the optimized Pareto solution set and transform it into an executable bill allocation instruction. This operation can be further implemented by selecting a compromise solution based on decision-maker preferences or by using the TOPSIS method to select the best solution from the Pareto set, thus outputting a final strategy with timeliness awareness, conflict coordination, and business adaptability.

[0075] Taking the centralized settlement of bills during peak periods for multiple business formats as an example, the automatic processing method for multi-business service bills in this embodiment can be as follows: In a certain month, a logistics company simultaneously processes three types of high-concurrency bills: cross-border, cold chain, and urban distribution. Cross-border bills mostly have a 30-day payment period, while urban distribution bills have a 7-day payment period. The system detects that the logistics business format scale is 3 and automatically increases the Pareto front resolution to distinguish subtle strategy differences. At the same time, based on the time discounting effect, the 7-day payment option is reset to 1.0, and the 30-day payment option is reduced to 0.7. During the optimization process, the algorithm identifies a strong conflict between minimizing costs and maximizing compliance. Therefore, it prioritizes compliance for high-weight urban distribution bills and allows moderate cost optimization for low-weight cross-border bills. Finally, the dynamic automatic processing strategy strictly enforces the service level billing rules for urban distribution bills and adopts a negotiated discount-based allocation ratio for cross-border bills, achieving a balance between overall financial stability and operational efficiency.

[0076] In one embodiment, the objective function includes one or more of the following: the objective function for minimizing the error rate, the objective function for minimizing bill processing time, and the objective function for maximizing settlement compliance.

[0077] The objective function for minimizing the error rate can be a mathematical expression aimed at minimizing the deviation between the billing result and the actual business attribution, used to measure financial accuracy. In an exemplary embodiment, the objective function for minimizing the error rate can drive the optimization process to prioritize avoiding billing errors caused by rule mismatch or data ambiguity. Furthermore, the objective function for minimizing the error rate can include, but is not limited to, one or more of the following: duplicate billing error items, omission error items, and cross-business attribution error items. The objective function for minimizing billing processing time can be a mathematical expression aimed at minimizing the time consumed from the original bill access to the billing strategy output. For example, the objective function for minimizing billing processing time can improve the system's responsiveness and throughput efficiency in high-frequency, real-time settlement scenarios. In a specific embodiment, the objective function for minimizing billing processing time can employ data parsing time items, matching calculation time items, and optimization solution time items.

[0078] The objective function for maximizing settlement compliance can be a mathematical expression aimed at maximizing the degree of compliance of the accounting logic with industry regulatory requirements and corporate internal control standards. In this embodiment, the objective function for maximizing settlement compliance ensures that the accounting results meet the compliance constraints of auditing, taxation, and internal risk control. Furthermore, the objective function for maximizing settlement compliance may include, but is not limited to, items related to compliance with financial and tax rules, consistency of contract terms, and integrity of data traceability. It may include one or more of the objective functions for minimizing error rates, minimizing billing processing time, and maximizing settlement compliance, which can be selected and combined according to the needs of the current business scenario when constructing the optimization objective function. Furthermore, one or more of the objective functions for minimizing error rates, minimizing billing processing time, and maximizing settlement compliance can be achieved by fixing the combination of the three objectives and configuring default weights, and dynamically enabling relevant sub-objectives based on the business context. This allows the optimization objectives to accurately correspond to actual business pain points and supports intelligent decision-making under multi-dimensional trade-offs.

[0079] Taking a mixed scenario of month-end centralized reconciliation and daily real-time settlement as an example, the automatic billing method for multi-business services in this embodiment can be as follows: During the month-end financial closing period, the system enables a dual objective function of minimizing the error rate and maximizing settlement compliance, tolerating a slightly longer processing time to ensure account accuracy and audit compliance; while in daily operations, for high-frequency, small-amount bills for urban delivery, the system only enables the objectives of minimizing bill processing time and minimizing the error rate, weakening the compliance weight to improve processing speed; both modes are executed through the same NSGA-III optimization framework, only the objective function composition is different, reflecting the system's flexible adaptability.

[0080] Furthermore, this embodiment of the invention also proposes a storage medium storing a multi-business service bill automatic processing program, which, when executed by a processor, implements the steps of the multi-business service bill automatic processing method described above.

[0081] In addition, refer to Figure 2 This invention also proposes an automatic billing system for multi-business services, which includes: Data acquisition module 10 is used to acquire original communication bill data of multi-business logistics and business rule parameters of each business; The matrix matching module 20 is used to match the original communication bill data of the multi-format logistics and the business rule parameters of each format, determine the matching matrix of the bill items to be split and the applicable scope of each format rule, and determine the split processing allocation strategy of the multi-format logistics service bill based on the matching matrix. The dynamic optimization module 30 is used to coordinate and optimize the accounting allocation strategy based on the optimization objective function of the accounting allocation strategy, and determine the dynamic automatic processing strategy.

[0082] Other embodiments or specific implementations of the multi-business service bill automatic processing system of the present invention can be referred to the above-described method embodiments, and will not be repeated here.

[0083] The above are merely preferred embodiments of the present invention and do not limit the scope of the patent. Any equivalent structural or procedural transformations made based on the description and drawings of the present invention, or direct or indirect applications in other related technical fields, are similarly included within the scope of patent protection of the present invention.

Claims

1. A method for automatically processing bills for multi-business services, characterized in that, The method includes: Obtain raw communication billing data for multi-business logistics and business rule parameters for each business; The original communication billing data of the multi-format logistics and the business rule parameters of each format are matched to determine the matching matrix of the billing items to be split and the applicable scope of each format rule, and the splitting processing allocation strategy of the multi-format logistics service bill is determined according to the matching matrix. Based on the optimization objective function of the revenue sharing allocation strategy, the revenue sharing allocation strategy is coordinated and optimized to determine the dynamic automatic processing strategy.

2. The method for automatically processing multi-business service bills as described in claim 1, characterized in that, The matching matrix used to determine the billing entries to be split and the applicable scope of rules for each business type includes: Based on the original communication billing data of the multi-business logistics, billing item features are extracted, and based on the business rule parameters of each business business, business rule adaptation features are extracted. Determine a first interaction relationship between the bill entry features and the business format rule adaptation features, and determine a coupling adaptation model between the bill entry and the business format rule based on the first interaction relationship; The coupling adaptation model is used to match the bill item features and the business rule adaptation features to determine the matching matrix between the bill items to be split and the applicable scope of each business rule.

3. The method for automatically processing multi-business service bills as described in claim 1, characterized in that, The step of determining the revenue sharing and allocation strategy for multi-format logistics service bills based on the matching matrix includes: The characteristic type of the bill item is determined based on the original communication bill data of the multi-business logistics; Determine the characteristic quantification indicators of the bill item feature types and the business rule parameters of each business format; Based on historical settlement data, cross-business shared bills and deduction items are used to identify abnormal adjustment and supplementary entries. The abnormal adjustment supplementary entry is used as the basis for supplementary adjustment of the disputed bill period. The supplementary adjustment basis and the feature quantification index are matched by the matching matrix to determine the billing and allocation strategy for multi-format logistics service bills.

4. The method for automatically processing multi-business service bills as described in claim 3, characterized in that, The determination of bill item feature types based on the original communication bill data of the multi-business logistics includes: Analyze the original communication billing data of the multi-business logistics to determine the second interaction relationship between the item attributes, amount scope, transaction scenario and business type attribution of the billing; Based on the second interaction relationship, determine the bill entry feature type.

5. The method for automatically processing multi-business service bills as described in claim 3, characterized in that, The characteristic quantification indicators for determining the feature type of the bill item and the business rule parameters of each business type include: Analyze the characteristic types of the bill items to determine quantitative indicators for bill attribution, quantitative indicators for bill time regularity, and quantitative indicators for bill amount fluctuation. Determine the third interaction relationship between the bill attribution quantitative indicator, the bill time regularity quantitative indicator, the bill amount fluctuation quantitative indicator, and the business rule parameters of each business format; The third interaction relationship is analyzed and quantified using cosine similarity to obtain characteristic quantitative indicators of business rules for each business format.

6. The method for automatically processing multi-business service bills as described in claim 3, characterized in that, The step of using the abnormal adjustment supplementary entry as the basis for supplementary adjustment of the disputed billing period, and matching the supplementary adjustment basis and the feature quantification indicators through the matching matrix to determine the billing allocation strategy for multi-format logistics service bills includes: Obtain quantifiable indicators of bill attribution, bill time regularity, and bill amount fluctuation; Based on the bill attribution quantitative indicator, the bill time regularity quantitative indicator, and the bill amount fluctuation quantitative indicator, the accounting rule adaptation parameters, abnormal threshold parameters, and trigger parameters for using abnormal adjustment supplementary entries as supplementary adjustments for disputed bills are determined for the automatic processing of bills for multi-business logistics services. Based on the revenue sharing rule adaptation parameters, the anomaly threshold parameters, and the triggering parameters, a revenue sharing processing and allocation strategy for multi-business logistics service bills is determined.

7. The method for automatically processing multi-business service bills as described in claim 1, characterized in that, The optimization objective function based on the revenue sharing allocation strategy coordinates and optimizes the revenue sharing allocation strategy to determine a dynamic automatic processing strategy, including: The Pareto front resolution is dynamically adjusted based on the adaptive reference point and the scale of the logistics industry using the third-generation non-dominated sorting genetic algorithm, and the weight of different accounts is determined based on the time discounting effect of the third-generation non-dominated sorting genetic algorithm. Determine the conflict mechanism among multiple objectives in the objective function; Based on the adjusted Pareto front resolution and the different account weights, the objective function is dynamically optimized by balancing the conflict mechanisms between the objectives and determining the dynamic automatic processing strategy for the account allocation strategy.

8. The method for automatically processing multi-business service bills as described in claim 1 or 7, characterized in that, The objective function includes one or more of the following: the objective function for minimizing the error rate, the objective function for minimizing bill processing time, and the objective function for maximizing settlement compliance.

9. A multi-business service billing automatic processing system, characterized in that, The multi-business service billing automatic processing system includes: The data acquisition module is used to acquire original communication billing data for multi-business logistics and business rule parameters for each business. The matrix matching module is used to match the original communication billing data of the multi-business logistics and the business rule parameters of each business, determine the matching matrix of the billing items to be split and the applicable scope of each business rule, and determine the splitting processing and allocation strategy of the multi-business logistics service bill based on the matching matrix. The dynamic optimization module is used to coordinate and optimize the revenue sharing allocation strategy based on the optimization objective function of the revenue sharing allocation strategy, and determine the dynamic automatic processing strategy.

10. A storage medium, characterized in that, The storage medium stores an automatic billing process for multi-business services, which, when executed by a processor, implements the steps of the automatic billing method for multi-business services as described in any one of claims 1 to 8.

Images