Development of a reliability based dynamic dredging decision model

Accurate forecases of hydrological and hydraulic conditions used in planning for the maintenance of navigation channels are important for ensuring safe and cost efficient water borne transportation. The uncertainties inherent in forecasting and in the availability and operating conditions of dredges make planning difficult and may require that an uncertainty assessment be integrated into the planning process for dredging activities. This paper develops a reliability based dynamic dredging decision model that employs a simulation-optimization approach combining a hydrological simulation model of stochastic channel conditions with a dynamic location model to schedule the optimal deployment and acitcity levels for dredges. The benefits of this type of approach are demonstrated by application to the problem of allocating different types and sizes of dredges in order to maintain a required channel depth over time and under different reliability levels. The amounts to be dredged at any given location are incorporated into the model as chance constraints that are dependant upon the stochastic hydrologic and hydraulic channel conditions. The multiobjective optimization model assigns demobilization and mobilization costs in periods when a facility is moved from one location to another, and allows for advanced maintenance dredging, ‘over-dredging’ in some time periods, in order to reduce overall costs. The dredging costs considered in the model are comprised of fixed costs-assessed each period and varying by size and type of dredge employed, variable costs-dependant upon the amount of material dredged in any month, and mobilization and demobilization costs-incurred each time a dredge is moved to a new location. The multiobjective framework evaluates explicitly the manner in which dredging costs vary under different levels of reliability for maintaining the required depth.