A Dynamic Multi-Agent Game Framework for Live Streaming E-Commerce Real-Time Comment Driven Virtual Anchor Discourse and Strategy Generation
Main Article Content
Abstract
Virtual anchors in live-streaming e-commerce often suffer from delayed responses to real-time comments, rigid preset scripts, and insufficient dynamic strategic reasoning. To address these limitations, this paper proposes a dynamic multi-agent game framework for real-time comment-driven virtual anchor discourse and strategy generation. First, a viewer state space is constructed by encoding bullet-screen comments through fine-grained sentiment and intent analysis. Second, a three-party dynamic Bayesian game is introduced among the virtual anchor, audience group, and competitor environment, and a deep Q-network is used to solve equilibrium policy labels in real time. Third, these strategy labels constrain a large language model to generate controllable, context-aware, and persona-consistent scripts. Finally, a dual-model verification mechanism enforces compliance and virtual-anchor identity consistency. Experimental results show that the proposed method increases conversion rate from 3.82% to 4.53% and extends average dwell time from 187.4 s to 238.3 s. The game-win rate reaches 64.9%, and script adaptation accuracy reaches 87.2%, confirming the effectiveness of game-theoretic modeling in real-time e-commerce interaction.
Downloads
Article Details
This work is licensed under a Creative Commons Attribution 4.0 International License.
Authors who publish with this journal agree to the following terms:
- Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under a Creative Commons Attribution License that allows others to share the work with an acknowledgement of the work's authorship and initial publication in this journal.
- Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgement of its initial publication in this journal.
- Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work (See The Effect of Open Access).
References
Y. Wang, Z. Lu, P. Cao, J. Chu, H. Wang, and R. Wattenhofer, “How Live Streaming Changes Shopping Decisions in E-commerce: A Study of Live Streaming Commerce,” Computer Supported Cooperative Work, vol. 31, no. 4, pp. 701-729, 2022, DOI: 10.1007/s10606-022-09439-2.
L. Li, Y. Feng, and A. Zhao, “An interaction-immersion model in live streaming commerce: the moderating role of streamer attractiveness,” Journal of Marketing Analytics, vol. 12, no. 3, pp. 701-716, 2024, DOI: 10.1057/s41270-023-00225-7.
N. Chen and Y. Yang, “The Role of Influencers in Live Streaming E-Commerce: Influencer Trust, Attachment, and Consumer Purchase Intention,” Journal of Theoretical and Applied Electronic Commerce Research, vol. 18, no. 3, pp. 1601-1618, 2023, DOI: 10.3390/jtaer18030081.
C. Kwan W, R. Wang H, M. Wang H, and F. Wong K, “A Survey on Recent Advances and Challenges in Reinforcement Learning Methods for Task-oriented Dialogue Policy Learning,” Machine Intelligence Research, vol. 20, no. 3, pp. 318-334, 2023, DOI: 10.1007/s11633-022-1347-y.
J. Wang, C. Zhang, D. Zhang, et al., “A recent survey on controllable text generation: A causal perspective,” Fundamental Research, vol. 5, no. 3, pp. 1194-1203, 2025, DOI: 10.1016/j.fmre.2024.01.001.
B. Zarouali, C. Boerman S, M. Voorveld H A, and G. van Noort, “The algorithmic persuasion framework in online communication: conceptualization and a future research agenda,” Internet Research, vol. 32, no. 4, pp. 1076-1096, 2022, DOI: 10.1108/INTR-01-2021-0049.
A. Braca and P. Dondio, “Persuasive communication systems: a machine learning approach to predict the effect of linguistic styles and persuasion techniques,” Journal of Systems and Information Technology, vol. 25, no. 2, pp. 160-191, 2023, DOI: 10.1108/JSIT-07-2022-0166.
H. Brinson N, L. Lemon L, C. Bender, et al., “Consumer response to podcast advertising: the interactive role of persuasion knowledge and parasocial relationships,” Journal of Consumer Marketing, vol. 40, no. 7, pp. 971-982, 2023, DOI: 10.1108/JCM-01-2023-5819.
J. Ni, T. Young, V. Pandelea, F. Xue, and E. Cambria, “Recent advances in deep learning based dialogue systems: a systematic survey,” Artificial Intelligence Review, vol. 56, no. 4, pp. 3055-3155, 2023, DOI: 10.1007/s10462-022-10248-8.
A. Algherairy and M. Ahmed, “A review of dialogue systems: current trends and future directions,” Neural Computing and Applications, vol. 36, no. 12, pp. 6325-6351, 2024, DOI: 10.1007/s00521-023-09322-1.