A State Space Framework for Coordination

这是一项为期一年的定期资助，研究智能代理社会中基于数学模型的协调方法。该研究解决了如何构建一个通用的协调模型，该模型基于对一组状态变量的识别，这些状态变量表征了代理社会中代理行为的相关方面。根据广泛使用的协调定义，协调被定义为对活动之间依赖关系的管理。协调的问题就是对这些依赖施加控制的问题。该研究旨在发现基于依赖关系的状态变量的逻辑顺序，从而建立协调策略的数学模型。一旦知道了协调和控制关系，就可以使用约束满足、约束传播和优化方法来处理协调问题。数学模型将通过仿真进一步探索和验证。

This is a one-year regular grant that investigates mathematical model-based methods for coordination in intelligent agent societies. The research addresses how to construct an general model of coordination that is based on a identification of a set of state variables that characterize relevant aspects of agent behavior in an agent society. Coordination is defined as the management of dependencies among activities, following a widely used definition of coordination. The problem of coordination is then the problem of exerting control over these dependencies. The research aims to discover logical orderings of state variables based on dependencies, and thus to develop mathematical models of coordination strategies. Once coordination and control relationships are known, coordination problems can be treated using constraint satisfaction, constraint-propagation, ad optimization approaches. The mathematical models will be further explored and validated through simulation.