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 2203.09655cs.GTcs.MA
收藏

与朋友、敌人和中立者的享乐游戏:解决开放性问题和细粒度复杂性

Hedonic Games With Friends, Enemies, and Neutrals: Resolving Open Questions and Fine-Grained Complexity

Jiehua Chen, Gergely Csáji, Sanjukta Roy, Sofia Simola

摘要:

我们研究了在有朋友、敌人和中立者的享乐博弈中突出稳定性概念的验证和存在性问题[8,16]。我们解决了几个(长期存在的)开放问题[4,16,20,23],并表明对于面向朋友的偏好,在朋友和敌人模型下,验证给定的代理分区是否(严格)核心稳定是NP完全的,而在朋友,敌人和中立者模型下,确定是否存在单个稳定分区是NP完全的。我们进一步研究了文献中的自然限制情况,例如当朋友和敌人的关系是对称的,当初始联盟有界的大小,当友谊图中的顶点度(分别是)。友谊和敌人图的并集)是有界的,或者当这样的图是非循环的或接近于非循环的时。我们获得了一个完整的(参数化)复杂性图片关于这些情况。

英文摘要:

We investigate verification and existence problems for prominent stability concepts in hedonic games with friends, enemies, and optionally with neutrals [8, 16]. We resolve several (long-standing) open questions [4, 16, 20, 23] and show that for friend-oriented preferences, under the friends and enemies model, it is coNP-complete to verify whether a given agent partition is (strictly) core stable, while under the friends, enemies, and neutrals model, it is NP-complete to determine whether an individual stable partition exists. We further look into natural restricted cases from the literature, such as when the friends and enemies relationships are symmetric, when the initial coalitions have bounded size, when the vertex degree in the friendship graph (resp. the union of friendship and enemy graph) is bounded, or when such graph is acyclic or close to being acyclic. We obtain a complete (parameterized) complexity picture regarding these cases.

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。

↑