Search

CN-121998737-A - Cross-border electronic commerce operation simulation system and method based on multi-agent simulation

CN121998737ACN 121998737 ACN121998737 ACN 121998737ACN-121998737-A

Abstract

The invention discloses a cross-border electronic commerce operation simulation system and method based on multi-agent simulation, which relate to the technical field of data processing and comprise the steps of executing cross-dimension segmentation according to collected transaction content data, interaction content data and cultural semantic data to obtain cross-cultural semantic data, and applying multidirectional mapping disturbance to the cross-cultural semantic data to generate behavior driving data; the method comprises the steps of acquiring policy and semantic data, acquiring the policy and semantic data, executing cross-structure splitting according to the acquired policy and semantic data to obtain policy constraint data, inputting the policy constraint data and behavior driving data into bidirectional entanglement calculation to generate preference evolution data, executing cultural domain difference based on the cultural semantic data and the behavior driving data to obtain differential behavior data, and executing cross-stitching calculation on the differential behavior data and the preference evolution data.

Inventors

  • GAO WEI

Assignees

  • 苏州沃金网络科技有限公司

Dates

Publication Date
20260508
Application Date
20260122

Claims (10)

  1. 1. A multi-agent simulation-based cross-border e-commerce operation simulation system, the system comprising: the segmentation generating module S11 is used for executing cross-dimension segmentation according to the collected transaction content data, the interaction content data and the culture semantic data to obtain cross-culture semantic data, and applying multidirectional mapping disturbance to the cross-culture semantic data to generate behavior driving data; The splitting entanglement module S12 is used for executing cross-structure splitting according to the collected policy and semantic data to obtain policy constraint data, inputting the policy constraint data and behavior driving data into bidirectional entanglement calculation, and generating preference evolution data; the differential stitching module S13 is used for implementing cultural domain difference based on cultural semantic data and behavior driving data to obtain differential behavior data, and executing cross stitching calculation on the differential behavior data and preference evolution data to obtain stitching preference data; And the coupling simulation module S14 is used for performing behavior coupling through the behavior driving data, the preference evolution data and the stitching preference data to obtain multi-agent updating data, and generating cross-border E-commerce operation simulation data based on the multi-agent updating data.
  2. 2. The multi-agent simulation-based cross-border e-commerce operation simulation method of claim 1, wherein the step of generating behavior-driven data comprises: S111, text segmentation is executed based on paragraph structure marks in transaction content data and interactive content data, and semantic segment index data are generated; S112, performing classification aggregation on the semantic segment index data based on the language tag index in the cultural semantic data to generate language dimension segment data; S113, performing semantic density measurement and classification boundary calculation based on the language dimension fragment data to generate cross-culture semantic data; S114, applying semantic disturbance function mapping to the cross-cultural semantic data to generate behavior driven data.
  3. 3. The multi-agent simulation-based cross-border e-commerce operation simulation method of claim 2, wherein the step of generating cross-cultural semantic data comprises: s113.1, constructing language fragment group data through fragment sets of the same language category in the language dimension fragment data; s113.2, performing semantic aggregation vector coding according to the language segment group data to generate preliminary language aggregation vector data; S113.3, calculating a semantic deviation value and a semantic overlap value based on the preliminary language aggregate vector data, and generating cross-culture semantic data.
  4. 4. The multi-agent simulation-based cross-border e-commerce operation simulation method of claim 3, wherein the step of generating preference evolution data comprises: S121, performing structure division based on the structure unit labels in the policy semantic data to generate structure block distribution data; S122, extracting constraint content segments based on high-frequency word segments in the structure block distribution data, and generating policy constraint data; S123, inputting policy constraint data and behavior driving data into a bidirectional entanglement calculation structure together to generate bidirectional action result data; s124, performing offset value evaluation based on the semantic reaction sites and the disturbance sensitive parameters in the bidirectional effect result data, and generating preference evolution data.
  5. 5. The multi-agent simulation-based cross-border e-commerce operation simulation method of claim 4, wherein the step of generating policy constraint data comprises: s122.1, extracting fragment content marked as restricted expression in the structure block distribution data, and constructing restricted fragment data; S122.2, performing language segment alignment processing on the restriction segment data, screening out repeated expression and semantic redundancy parts, and generating simplified constraint content data; S122.3, inputting the simplified constraint content data into a constraint information screening function to generate policy constraint data.
  6. 6. The multi-agent simulation-based cross-border e-commerce operation simulation method of claim 5, wherein the step of generating stitching preference data comprises: S131, constructing a semantic deviation comparison matrix based on cultural semantic data and behavior driving data, and generating cultural domain difference data; s132, performing difference combination processing on the cultural domain difference data and the preference evolution data to generate difference behavior data; S133, generating stitching preference data based on the segment-level variation range in the differential behavior data to match the offset section in the preference evolution data.
  7. 7. The multi-agent simulation-based cross-border e-commerce operation simulation method of claim 6, wherein the logic for generating the differential behavior data comprises: s132.1, extracting high-offset fragment data in cultural domain difference data, and constructing a behavior difference fragment set; S132.2, performing difference fitting analysis on the behavior difference fragment set and the corresponding paragraph in the preference evolution data to generate fragment-level behavior difference data; and S132.3, performing semantic concentration coding processing based on the fragment-level behavior difference data to generate differential behavior data.
  8. 8. The multi-agent simulation-based cross-border e-commerce operation simulation method as claimed in claim 7, wherein the step of generating cross-border e-commerce operation simulation data comprises: S141, constructing a behavior preference combination mapping diagram based on behavior driving data and preference evolution data, and generating combination behavior expression data; s142, performing agent state mapping based on the stitching preference data and the combined behavior expression data to generate multi-agent update data; And S143, performing behavior sequence simulation based on the multi-agent update data to generate cross-border e-commerce operation simulation data.
  9. 9. The multi-agent simulation-based cross-border e-commerce operation simulation method of claim 8, wherein the logic for generating the multi-agent update data is: S142.1, extracting a behavior consistency segment from the stitching preference data and the combined behavior expression data, and constructing combined input data; S142.2, inputting the joint input data into a behavior coupling function, and executing agent preference splitting processing to generate agent deviation difference data; and S142.3, performing update instruction allocation and multi-agent structure recombination based on the agent deviation difference data to generate multi-agent update data.
  10. 10. A multi-agent simulation-based cross-border e-commerce operation simulation method applied to the multi-agent simulation-based cross-border e-commerce operation simulation system as claimed in any one of claims 1 to 9, characterized in that the method comprises: S21, executing cross-dimension segmentation according to the collected transaction content data, interaction content data and cultural semantic data to obtain cross-cultural semantic data, and applying multidirectional mapping disturbance to the cross-cultural semantic data to generate behavior driving data; S22, performing cross-structure splitting according to the collected policy and semantic data to obtain policy constraint data, inputting the policy constraint data and behavior driving data into bidirectional entanglement calculation, and generating preference evolution data; s23, implementing cultural domain difference based on cultural semantic data and behavior driving data to obtain differential behavior data, and executing cross stitching calculation on the differential behavior data and preference evolution data to obtain stitching preference data; And S24, performing behavior coupling through the behavior driving data, the preference evolution data and the stitching preference data to obtain multi-agent updating data, and generating cross-border electronic business operation simulation data based on the multi-agent updating data.

Description

Cross-border electronic commerce operation simulation system and method based on multi-agent simulation Technical Field The invention relates to the technical field of data processing, in particular to a cross-border electronic commerce operation simulation system and method based on multi-agent simulation. Background In the multi-cultural environment-oriented cross-border electronic commerce operation modeling process, the problem that a user portrait template is highly dependent generally exists, a user behavior simulation effect is limited by template precision and application range, when cultural semantics are fuzzy in boundary or drifting in context, a modeling result often cannot accurately reflect behavior change caused by cultural differences, meanwhile, a consumer behavior decision process often presents a recessive evolution trend, the traditional modeling method is difficult to effectively respond to the dynamic change through static tags or predefined features, so that the generalization capability of behavior simulation is insufficient, and the consumption mode reconstruction and intelligent body evolution under a heterogeneous cultural background are difficult to support. In summary, the technical problem to be solved is how to realize the behavior simulation and the preference dynamic evolution of the consumer intelligent body under the multi-culture multi-policy background under the conditions of not depending on the prior user portrait template and not sacrificing the modeling calculation efficiency. Disclosure of Invention In order to solve the technical problems, the invention provides a cross-border electronic commerce operation simulation system and method based on multi-agent simulation. A multi-agent simulation-based cross-border e-commerce operation simulation system, the system comprising: the segmentation generating module S11 is used for executing cross-dimension segmentation according to the collected transaction content data, the interaction content data and the culture semantic data to obtain cross-culture semantic data, and applying multidirectional mapping disturbance to the cross-culture semantic data to generate behavior driving data; The splitting entanglement module S12 is used for executing cross-structure splitting according to the collected policy and semantic data to obtain policy constraint data, inputting the policy constraint data and behavior driving data into bidirectional entanglement calculation, and generating preference evolution data; the differential stitching module S13 is used for implementing cultural domain difference based on cultural semantic data and behavior driving data to obtain differential behavior data, and executing cross stitching calculation on the differential behavior data and preference evolution data to obtain stitching preference data; And the coupling simulation module S14 is used for performing behavior coupling through the behavior driving data, the preference evolution data and the stitching preference data to obtain multi-agent updating data, and generating cross-border E-commerce operation simulation data based on the multi-agent updating data. Further, the step of generating behavior-driven data includes: S111, text segmentation is executed based on paragraph structure marks in transaction content data and interactive content data, and semantic segment index data are generated; S112, performing classification aggregation on the semantic segment index data based on the language tag index in the cultural semantic data to generate language dimension segment data; S113, performing semantic density measurement and classification boundary calculation based on the language dimension fragment data to generate cross-culture semantic data; S114, applying semantic disturbance function mapping to the cross-cultural semantic data to generate behavior driven data. Further, the step of generating cross-cultural semantic data includes: s113.1, constructing language fragment group data through fragment sets of the same language category in the language dimension fragment data; s113.2, performing semantic aggregation vector coding according to the language segment group data to generate preliminary language aggregation vector data; S113.3, calculating a semantic deviation value and a semantic overlap value based on the preliminary language aggregate vector data, and generating cross-culture semantic data. Further, the step of generating preference evolution data includes: S121, performing structure division based on the structure unit labels in the policy semantic data to generate structure block distribution data; S122, extracting constraint content segments based on high-frequency word segments in the structure block distribution data, and generating policy constraint data; S123, inputting policy constraint data and behavior driving data into a bidirectional entanglement calculation structure together to generate bidirectional action result data