arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2411.03335cs.SI
收藏

非对称加权级联模型用于竞争性影响力最大化

Asymmetric Weighted Cascade Model for Competitive Influence Maximization

Vipin Gunda, Archit Mehta

AI总结:

本文提出非对称加权级联模型,研究竞争性影响力最大化问题,证明其NP难,并发现预算与产品评分在不同网络直径下的优势差异及图规模与影响节点数的线性关系。

摘要:

我们引入了一种改进的加权级联模型,该模型整合了非对称预算和产品评分,为广义非对称影响力最大化问题提供了新的见解,我们证明该问题是NP难的。我们的模拟表明,在直径较大的网络中,预算较高的参与者具有明显优势,而在直径较小的网络中,产品评分较高的参与者则表现出显著优势。此外,我们确定了图规模与被影响节点数量之间存在稳健的线性关系。在密集连接的网络中,我们推导出了独立于网络规模的影响力概率界限。我们对该领域中纳什均衡的考察强调了不存在保证的纯纳什均衡,这表明在策略上提升预算或产品评分可能比在此背景下追求最优策略带来更实质性的收益。

英文摘要:

We introduce a modified Weighted Cascade model that integrates asymmetric budgets and product scores, providing new insights into the Generalized Asymmetric Influence Maximization problem, which we establish as NP-hard. Our simulations demonstrate that players with higher budgets possess a distinct advantage in networks characterized by larger diameters, whereas players with superior product scores exhibit a significant advantage in networks with smaller diameters. Moreover, we identify a robust linear relationship between graph size and the magnitude of influenced nodes. In densely connected networks we derive bounds for the probabilities of influence that are independent of network size. Our examination of Nash equilibria in this domain underscores the absence of a guaranteed pure Nash equilibrium, suggesting that the strategic enhancement of budgets or product scores may yield more substantial benefits than the pursuit of an optimal strategy in this context.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑