Feedback method of control for estuarine management

The purpose of this study is to develop an efficient procedure for computing the optimal freshwater inflows into bays and estuaries while satisfying the freshwater requirements for other components in the system. The estuary management model combines a linear transition function derived from the salinity regression equations (which is a function of fresh water inflow) with an optimal control algorithm. Differential dynamic programming (DDP), and the feedback method of control as applied to discrete-time, constrained optimal control problems are used.; The constraints of the model are, the monthly freshwater inflows, monthly salinity, mean monthly inflows in seasons and the annual fish harvest of various species. The constrained DDP and feedback method of control employs a penalty function of the form, P {dollar}sp*{dollar} min {dollar}lbrack{dollar}0,g(x){dollar}rbracksp2{dollar}, where P is the penalty weight, g(x) is the constraint, and x contains the design variables. A modification of DDP by Li, in 1995 and the feedback method of control which makes these techniques applicable to estuary management problem is discussed. The modified DDP and the feedback method of control are applied to the Lavaca-Tres Palacios Estuary in Texas.; In the application of DDP, the non convexity of the objective function is overcome by applying a shift method developed by Liao and Shoemaker, in 1991. Also gaussian elimination, LU decomposition and singular value decomposition are used to compute the inverse matrices in the DDP algorithm. These different alternatives are compared in terms of accuracy and convergence rate. The estuary management problem is solved by the nonlinear programming code GAMS-MINOS as well. The similarity of the optimal results for DDP and the feedback method of control with GAMS-MINOS guarantee the optimality of the estuary management model.; In the application of the modified feedback method, the penalty function technique is modified and a shift procedure is not required to obtain feasible solutions. In order to increase the accuracy of the optimal results a hydrodynamic transport model must replace the regression equations. The hydrodynamic salinity model, HYD-SAL, is analyzed in detail for the purpose of interfacing it with DDP and the feedback method of control.