Grid computing provides key infrastructure for distributed problem solving in dynamic virtual organizations. Many scientific projects have adopted grid computing, and industrial interest in it is rising rapidly. However, grids are still the domain of a few highly trained programmers with expertise in networking, high-performance computing, and operating systems. We aim to capture knowledge and heuristics about how to select application components and computing resources in a planning-based system, to help generate executable job workflows for a grid. We?ve integrated the planner with a grid environment, where it has generated in real time dozens of workflows with hundreds of jobs. To be applicable to a wide range of existing and new grid applications, the planner provides different levels of support, depending on the information available.