A number of investigations have been performed in the recent years to analyze the effect of advance maintenance and to see if it is justifiable. These studies gave promising results but were site specific and considered only a short period of shoaling after dredging. Lund considered the above problem as a least-cost replacement problem and determined the optimal dredging cycle for a single reach with a constant shoaling rate. Due to uncertainty in the shoaling rate and the river flow, the scheduling of advance maintenance (AM) is a probabilistic sequential decision process. This paper discusses the application of a modified policy iteration algorithm developed by Tjims to determine the least-cost inspection and maintenance schedule. The method takes into account the dredging cost, cost incurred due to failure, and the cost for inspection in determining an average cost optimal policy.