Abstract Nowadays, it is possible to easily utilize positive and negative effects of neighbors on a social network to maximize diffusion of a novel product and profit of the seller. Hence, this paper aims to introduce a new mathematical model for a product pricing in non-competitive environment having multiple goals. The proposed model is designed while there are a monopole seller and several heterogeneous customers for a novel product. Considering various criteria, these customers are able to purchase the novel product including price, product quality, urgent need to have the product, and positive/negative externalities received from the neighbors. Moreover, they are able to comment in case of satisfaction or dissatisfaction with the product. However, the extent of influence depends on strength of the relations with neighbors that is considered in the proposed model with complete information and quantitative values. Proportionate to activating the neighbors, referral bonus is considered from the seller. To find influential nodes for the influence and exploit strategy implementation we propose a new overlapping community detection algorithm. In this algorithm, a new overlapping score based on non-member neighbor nodes connectivity is introduced to identify overlapping communities. Finally, we evaluate the efficiency of the proposed model, by implementing the proposed community detection algorithm in a real-world dataset. The results show that it is possible to obtain desired selling price in a fashion that maximum diffusion in the network happens and the seller achieves his desired profit under various management viewpoints.

中文翻译：

---

基于重叠社区关键节点识别的单层加权网络多目标优化模型

摘要 如今，可以很容易地利用邻居对社交网络的正面和负面影响来最大化新产品的传播和卖家的利润。因此，本文旨在为具有多个目标的非竞争环境中的产品定价引入一种新的数学模型。所提出的模型是在一个新产品有一个单极卖家和几个异构客户时设计的。考虑到各种标准，这些客户能够购买新产品，包括价格、产品质量、对产品的迫切需要以及从邻居那里收到的正/负外部性。此外，他们可以在对产品满意或不满意的情况下发表评论。然而，影响的程度取决于在具有完整信息和定量值的建议模型中考虑的与邻居关系的强度。与激活邻居成比例，卖家会考虑推荐奖金。为了找到影响和利用策略实施的有影响力的节点，我们提出了一种新的重叠社区检测算法。在该算法中，引入了基于非成员邻居节点连通性的新重叠分数来识别重叠社区。最后，我们通过在真实世界数据集中实施所提出的社区检测算法来评估所提出模型的效率。