Advance Reservation (AR) of resources in grid systems allows users to secure resources prior to executing their jobs. This is very important, specially for time critical and workflow applications. One major challenge for applications having interdependent tasks is to minimise the delay of execution of the whole application. In general, the problem of mapping a set of interdependent tasks on distributed services is a NP-complete problem. Thus, in practice, heuristics are most often used to schedule workflow applications in grid. We have exploited some important properties like slack time, critical path of workflow applications to provide a scheduling algorithm that reduces the average delay to complete execution of an application, increases the number of zero delayed applications and takes less average time to run. The system has been designed to support advance reservation of resources. Our proposed system is more fault tolerant and addresses this problem efficiently.