Grid computing infrastructures embody a cost-effective computing paradigm that virtualises heterogenous system resources to meet the dynamic needs of critical business and scientific applications. These applications range from batch processes and long-running tasks to more real-time and even transactional applications. Grid schedulers aim to make efficient use of Grid resources in a cost-effective way, while satisfying the Quality-of-Service requirements of the applications. Scheduling in such a large-scale, dynamic and distributed environment is a complex undertaking. In this paper, we propose an approach to Grid scheduling which abstracts over the details of individual applications and aims to provide a globally optimal schedule, while having the ability to dynamically adjust to varying workload demands using various capacity planning techniques. Our model places particular emphasis on the stochastic and upredictable nature of the Grid, leading to a more accurate reflection of the state of the Grid and hence more efficient and accurate scheduling decisions.