Grid technologies enable large-scale resource sharing among enterprises and/or individuals. In such settings, control policies must be administered so that resource can be fairly shared among participating enterprises. A grid-based clearinghouse for managing shared resources is built with the aim of promoting technical and scientific cooperation within and between campuses. With that, we need a global mechanism for exchanging and integrating resource and hence, maximize computation throughput through the coordination of shared resources. The major process involved is the development of a job management strategy. In this paper, we will introduce the design and application of a decentralized hierarchical scheduler for job management in the clearinghouse.