Meta-scheduling of level-set methods in a grid computing environment

Many complex algorithms demand computational resources that exceed those available in a single computer, with that demand continually growing to meet the advancement in computing resources. If leveraged efficiently, distributed computing environments may be able provide the computational resources needed to solve these problems. However, the adaptation of algorithms capable of leveraging such distributed environments remains a challenge that requires specialized expertise. The results of the work contained in this thesis provide a mechanism for use in the adaptation of a given problem for execution in a grid computing environment.;Grid technologies provide a layer of abstraction that lies between heterogeneous, distributed physical computational resources and the user of the resources. A class of problems that relies on the level-set method will be addressed throughout this thesis as the problem being adapted to the grid environment. These methods are used in the implicit representation of surfaces, support a wide array of surface motions, and allow evolution of interfaces as zero level-set functions of a time-dependent scalar field.;The simulation of grid resources allows analysis to be done against the various variables in the grid computing environment and problem domain space without introducing the need for a large-scale grid environment on which to execute various test cases. The use of these resources for testing is not an effective use of those resources, given the value one can gain through the use of simulation techniques. The Open Science Grid (OSG) is used as a real-world grid environment with computational resources that can be modeled in the simulation environment and allow for real-life realization of the predictions and appropriate performance indicators found through simulation.;This primary result in the thesis demonstrates the use of a simulation environment that models the OSG and the level-set method problem to provide an understanding of how to decompose the level-set algorithms for grid execution and what grid meta-scheduling paradigm is best suited for that execution on the OSG. A secondary result shows how to abstract the technique used to develop that result as a framework that can be leveraged for problems from other application domains and for execution on other target grid environments.