In this paper we address the problem of static scheduling of real-time systems that include both hard and soft tasks. We consider that hard as well as soft tasks are periodic and that there exist data dependencies among tasks. In order to capture the relative importance of soft tasks and how the quality of results is affected when missing a soft deadline, we use utility functions associated to soft tasks. Thus our objective is to find an execution order for tasks that maximizes the total utility and at the same time guarantees hard deadlines. We use the expected duration of tasks for evaluating utility functions whereas we use the maximum duration of tasks for ensuring that hard deadlines are always met. We present an algorithm for finding the optimal schedule and also different heuristics that find near-optimal solutions at reasonable computational cost. The proposed algorithms are evaluated using a large number of synthetic examples.