Collaborative Research: MW: Master-Worker Middleware for Grids

该项目将部署一个强大、可靠、易于使用的工具，使复杂的数值算法能够在网格上执行。这个项目的重点将是部署MW，这是一个软件工具，封装了主工作范式的抽象，作为NMI软件发布的一部分。在数值优化的几个重要的问题类别中，利用微单元来实现新的算法。这项工作将使MW成为生产质量的工具，增强其易用性和健壮性，并将其与其他NMI中间件集成。目标是使MW成为网格上主工中间件的事实上的标准。该建议的智力价值在于中间件的部署，它促进了在计算网格上解决科学和工程应用的各种复杂数值算法的实际实现。该工具将为NMI增加一个重要的维度。提供一个全面的、易于使用的工具，使研究人员、实践者和教育者能够利用计算网格的力量，其广泛的影响可能是深远的。研究人员可以模拟更复杂的物理现象模型。从业者可以解决更大的问题，更准确地接近所考虑的系统。教育工作者可以通过演示和模拟复杂的系统，为学生带来“复杂的课堂”。

This project will deploy a robust, reliable, easy-to-use tool that enables complex numerical algorithms to be executed on Grids. The thrust of this project will be to deploy MW, a software tool that encapsulates the abstractions of the master-worker paradigm, as part of the NMI software release. MW is used to implement new algorithms for several important problem classes in numerical optimization. The work will make MW a production-quality tool, enhance its ease of use and its robustness, and integrate it with other NMI middleware. The goal is to make MW the de facto standard for master-worker middleware on Grids.The intellectual merit of this proposal lies in the deployment of middleware that facilitates practical implementation of a wide range of complex numerical algorithms for solving science and engineering applications on computational Grids. The tool will add an important dimension to NMI. The broader impact of providing a comprehensive, easy-to-use tool allowing researchers, practitioners, and educators to harness the power of the computational Grid is potentially far-reaching. Researchers can simulate more complex models of physical phenomena. Practitioners can solve larger problems that more accurately approximate the system under consideration. Educators can bring the "complex to the classroom" by demonstrating and simulating complex systems for the benefit of their students.