Abstract:

This paper studies the impact of the ‘virus’ marketing strategy on the classical newsvendor problem. In the market, the relationships between all consumers form a so called ‘small-world’ social network. Therefore, the firm proposes to startup a virus marketing practice on this social network. Besides, to find a source node to diffuse word-of-mouth, the firm also needs to make decision on the order quantity to maximize the profit. The simulation results show that the solution based on mathematics analysis overestimates the demand and leads to a serious loss of profit. More importantly, the results reveal that the firm could benefit efficiently from tighter relationships, better acceptance among consumers, and the decrease in unsold cost of unit product. In addition, closeness centrality by social network analysis is a better method to choose a source node for virus marketing on small-world network.

Key words: newsvendor problem, virus marketing, small-world network, multi-agent modeling and simulation, social network

摘要： 以报童问题为基准问题，采用多智能体建模方法分析当企业实施病毒营销后的产品订购量问题。鉴于市场中消费者组成的“小世界”社会关系网络，企业将从所有消费者中选择一个作为源头节点，实施其病毒营销策略。同时，企业需要根据病毒营销的结果设定一个收益最大的产品订货量。通过模型的仿真求解，验证了此类问题数学分析的局限性。仿真结果表明：企业可以通过增强用户群体之间的关系、用户的口碑接受程度以及降低产品单位成本来有效提高收益。并且，社会网络分析中接近中心性指标是用于选择病毒营销源头节点的最佳评价指标。

关键词: 报童问题, 病毒营销, 小世界网络, 多智能体建模与仿真, 社会关系网络