تدوين مدل بهينه بهره داري از سيستم هاي چند مخزني با استفاده از روش كريدور در DP

Memory and computer time consumption are the two major obstacles in applying dynamic programming to complex multi-reservoir systems. This challenge has initiated the so-called curse of dimensionality solutions in water resources studies. Among many others, Discrete Differential Dynamic Programming (DDDP) is a mean to deal with the dimensionality problem. The method, which is based on the traditional dynamic programming, narrows the solution space through the assumption of a corridor around a nominal trajectory. Optimization is carried out through the iterations around this trajectory. At the end of each iteration, the final solution is assumed as the new nominal trajectory for the next iteration. The process is continued until no further improvement is possible. In this paper, a general multi-purpose multi-reservoir computer program is developed that is flexible in accepting any type of system configuration. The proposed model, in addition of using DDDP method, can handle joint water supply allocations with reservoirs in series or parallel. It is also possible to reduce the firm demand deficits, such as municipal needs, through water transfer from upstream allocations to non-firm uses, such as agricultural demands. The model was applied to a complex system with seven reservoirs and several demand points. The results indicate a distinguished improvement over the usual DP method. It was noticed that the appropriate assumption of state variable and initial nominal trajectory are crucial in reaching the optimum solution. Furthermore, changing the number of states has an insignificant effect on finding the final optimum solution. It should be noted that, there are no restrictions on the number of reservoirs within the model; however, available CPU and time consumption must be considered in modeling complex systems.