Mobility, Search and Communication

当今移动代理计算中一些最引人注目的发展是由于通信、搜索和移动性之间的协同作用而发生的。移动机器人配备了运动机制，并配备了感知硬件和软件，可以移动和感知世界。移动机器人需要通过控制其运动并以最有效和高效的方式应用其感知能力来在明确定义的领域中导航，同时将其发现传达给其他机器人，以促进明确定义的网络任务的解决。目前对协作移动代理(机器人、传感器)系统的研究缺乏对移动性在执行基本通信和搜索任务中的影响的了解。更具体地说，我们有兴趣了解以下问题。一群移动代理如何在分布的互联环境中高效地导航以完成其计算任务？可以利用哪些代理功能和哪些知识？一个特定问题有什么特别之处，使得它可以有效地解决，而不是替代方案？ 该提案将在移动代理计算的背景下追求三个互补的研究主题。1)当使用代理移动性来优化各种网络任务时，如何产生关键机制？2)当所有任务都由所有代理完成时，在图和几何域中进行最优合作搜索的代价是多少？3)交互图如何影响网络中种群协议的稳定性和收敛？本项目旨在通过追求这三个研究方向来增进我们的知识。提出的课题有望在理论计算机科学和分布式计算的发展中发挥重要作用。我们的方法将在分布式计算的背景下使用算法设计方法以及概率和确定性技术。 我们方法的一般指导原则如下。1)基于算法分布式计算的方法，提供对动态网络内交互通信实体的行为的更深层次的理解。2)使用严格的数学模型开发研究技术，这些模型满足严格和明确定义的要求，并出于对无线自组织和传感器网络、机器人搜索和移动、生物分布式计算和社会网络等已建立的研究空间的实际考虑。3)培养高素质的PDF、博士和硕士人才以及这些新兴研究领域的有天赋的本科生。

Some of the most striking developments in mobile agent computing today are occurring due to the synergy between communication, search and mobility. Equipped with locomotion mechanisms and outfitted with hardware and software for perception, mobile robots can move and perceive the world. A mobile robot is required to navigate through a well-defined domain by controlling its motion and applying its perception capabilities in the most effective and efficient manner, while at the same time communicating its findings to the rest of the robots so as to facilitate the solution of well-defined network tasks. Current research on cooperative mobile agent (robot, sensor) systems suffers from a lack of understanding of the impact of mobility in performing fundamental communication and search tasks. More specifically, we are interested in understanding the following questions. How can a swarm of mobile agents navigate efficiently through a distributed interconnected environment to complete its computing tasks? Which agent capabilities and what knowledge can be leveraged? What is special about a particular problem that makes it solvable efficiently as opposed to alternatives? This proposal will pursue three complementary research themes within the context of mobile agent computing. 1) How do critical regimes arise when agent mobility is employed to optimize various network tasks? 2) What is the cost of optimal co-operative search in graphs and geometric domains when all tasks are finished by all agents? 3) How do interactive graphs affect the stability and convergence of population protocols in networks? This project aims to advance our knowledge by pursuing these three research directions. Topics proposed are expected to play a significant role in the development of theoretical computer science and distributed computing. Our approach will employ algorithmic design methodologies as well as probabilistic and deterministic techniques in the context of distributed computing. The general guiding principles of our approach are the following. 1) Provide a deeper understanding of the behavior of interacting communicating entities within a dynamic network based on the methodologies of algorithmic distributed computing. 2) Develop research techniques using rigorous mathematical models that satisfy strict and well-defined requirements and are motivated from practical considerations in the established research space of wireless ad hoc & sensor networks, robot search & mobility, biological distributed computing, and social networks. 3) Train Highly Qualifies Personnel at the PDF, PhD and Master's level as well as gifted undergraduates in these emerging research fields.