Optimal Reservoir Rule Curves Considering Conditional Ant Colony Optimization with Simulation Model

Reservoir rule curves are guideline for long term operation of multi-purpose reservoir that affected from severe flood and drought situations. This study proposed an alternative technique for searching optimal rule curves of multi-purpose reservoir. The proposed model consists of a Conditional Ant Colony Optimization (CACO) and a reservoir simulation model. Monthly rule curves of the Lampao Reservoir located in the northeast of Thailand was considered in this study. Four hundred samples of generated inflow data of reservoir were used to evaluate the performance of the proposed model and the results were compared with those of the genetic algorithm technique (GAs) and currently used. The results found that the new rule curves produced by the CACO and simulation method provided a similar pattern of rule curves as compared with the rule curves provided by genetic algorithm technique. The patterns are different from the existing pattern. The situations of water shortage and flood of using the new rule curves with reservoir operation considering generated inflow are lower than those of using existing rule curves. However, these situations are closely to the situations of using the rule curves of genetic algorithms technique. In conclusion, the proposed model could enhance the performance of the Lampao Reservoirs and it might be applied to other reservoirs by modifying the objective functions and constraint equations of searching process.