Abstract:

Empirical studies show that influences among consumers are quite different when potential consumers are making their purchase decision. However, heterogeneous influences were usually simplified as a binary variable in analytical models, which made the analysis results obviously inconsistent with the empirical data. Therefore, this paper, regarding the heterogeneous influences as a directed, weighted social network, formulated demand as the final results of viral marketing strategies in this directed, weighted social network. Besides, facing the stochastic demand, it calculated the optimal order quantity with loss-averse decision-maker. Moreover, in order to mimic the process of viral marketing on the social network, it introduced multi-agent based modeling and simulation, in which each consumer was modeled as an autonomous agent, whose purchasing decision was influenced by his or her neighbors. Furthermore, it described viral marketing as an adapted SIR model. Finally, based on the simulation data and analysis, it proposed suggestions on the impact of the source node for viral marketing strategies, loss-averse model on the optimal order quantity and the profit of the firm.

Key words: viral marketing, loss-averse, newsvendor problem, weighted network, multi-agent modeling and simulation

摘要：

现实中，消费者之间的产品购买决策影响差异性较强，但是理论研究通常过度简化了这种相互影响的差异性。鉴于此现状，以一个有向加权的无标度网络模拟消费者之间的社会关系网络，分析企业病毒式营销策略下的产品市场需求。考虑到市场需求的不确定性，研究了损失规避风险偏好的企业决策者的最优订货量决策问题。通过多智能体建模与仿真，对企业病毒式营销策路下的源头节点选择、不同风险偏好及模型对企业最优订货量决策给出了合理的数据解释和运营建议。

关键词: 病毒式营销, 损失规避, 报童问题, 加权网络, 多智能体建模与仿真