arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1012.3282cs.GT
收藏

激励博弈与风险管理机制

Incentive Games and Mechanisms for Risk Management

Tansu Alpcan

摘要:

激励在具有多个自治部门的大型组织的(安全和IT)风险管理中发挥着重要作用。本文提出了一个基于博弈论的风险管理激励机制设计框架。风险管理者作为一个机制设计者,提供规则和激励因素,如援助或补贴的部门或单位,这是建模为自私的球员的战略(非合作)游戏。在此基础上,提出了满足效率、偏好相容性和防策略性准则的多目标激励机制。此外,迭代和分布式算法,它可以实现下的信息限制,如风险管理人员不知道个别单位的喜好。一个示例场景用数字说明了框架和结果。提出的激励机制设计方法不仅对制定指导方针,而且对开发大规模风险管理的计算机辅助系统也是有用的。

英文摘要:

Incentives play an important role in (security and IT) risk management of a large-scale organization with multiple autonomous divisions. This paper presents an incentive mechanism design framework for risk management based on a game-theoretic approach. The risk manager acts as a mechanism designer providing rules and incentive factors such as assistance or subsidies to divisions or units, which are modeled as selfish players of a strategic (noncooperative) game. Based on this model, incentive mechanisms with various objectives are developed that satisfy efficiency, preference-compatibility, and strategy-proofness criteria. In addition, iterative and distributed algorithms are presented, which can be implemented under information limitations such as the risk manager not knowing the individual units' preferences. An example scenario illustrates the framework and results numerically. The incentive mechanism design approach presented is useful for not only deriving guidelines but also developing computer-assistance systems for large-scale risk management.

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。

↑