Optimizing the production of electrical energy through the efficient management of reservoirs is one of the major concerns in power supply. Then, optimization methods have been employed to solve the problem of service continuity in a generation. Many reservoir management issues are indeed addressed without achieving a desired solution in line with demand. The main motivation for state-of-the-art technology is to develop robust and intelligent methods that enable strict compliance with production restrictions while meeting energy demand. This paper follows this perspective by proposing the water supply and hydroelectric production through a hybrid method combining the differential evolution and two constraints processing treatments (TCPT). The interest of the constraint method using here, including its superiority of feasible solutions (SF) and auto-adaptative penalty (AP) is its ability to overcome premature convergence and to obtain very accuracy near optimal solutions. AP technique improve by its ability process through tedious trial and error methods to select the best penalty coefficient during implementation. SF, for its part, is a technique for comparing unrealistic solutions on the basis of the violation of global constraint aimed at pushing unrealistic solutions towards the feasible region, while comparing two workable solutions on objective value improves the overall solution. To compare the performance of the proposed hybrid method of optimization with the SF and AP techniques as a constraint method, a series of experiments was firstly carried out using 25 well know benchmark functions. The obtained results as opposed to the current hybrid method such as TCPT-PSO, TCPT-ABC and TCPT-GA have exhibited the method’s superiority. Based on these performances, we employed the proposed hybrid method to manage the water supply and hydroelectric production of the Memve’ele reservoir. The application on the management of Memve’ele Reservoir demonstrates its accuracy in providing a near-optimal solution on the order of 99.7% for 1945s, 1985s and 1988s of demand during the seven hydrological years of the Ntem Valley with real mean square error (RMSE) in order of for water supply and hundredth for hydropower reservoir operation.

中文翻译：

---

水电生产优化管理：Memve'ele 水电站水库政策案例

通过水库的有效管理来优化电能生产是电力供应的主要问题之一。然后，采用优化方法来解决一代中的服务连续性问题。许多水库管理问题确实得到了解决，但没有达到符合需求的理想解决方案。最先进技术的主要动机是开发稳健且智能的方法，在满足能源需求的同时严格遵守生产限制。本文遵循这一观点，提出通过结合差分进化和两个约束处理处理（TCPT）的混合方法来供水和水力发电。这里使用的约束方法的优点，包括其可行解（SF）和自适应惩罚（AP）的优越性，在于它能够克服早熟收敛并获得接近最优解的高精度。 AP技术的改进在于其在实施过程中通过繁琐的试错方法选择最佳惩罚系数的能力。 SF 是一种在违反全局约束的基础上比较不切实际的解决方案的技术，旨在将不切实际的解决方案推向可行区域，同时比较两个可行的解决方案的客观价值，从而改善整体解决方案。为了比较所提出的混合优化方法与 SF 和 AP 技术作为约束方法的性能，首先使用 25 个众所周知的基准函数进行了一系列实验。所获得的结果相对于当前的混合方法如TCPT-PSO、TCPT-ABC和TCPT-GA显示了该方法的优越性。基于这些性能，我们采用了所提出的混合方法来管理 Memve'ele 水库的供水和水力发电。 Memve'ele 水库管理中的应用表明，它能够为 Ntem 山谷 7 个水文年期间 1945 年、1985 年和 1988 年的需求提供接近 99.7% 的近乎最优解决方案，且具有真实均方误差 (RMSE) ) 供水为第 100 位，水电站水库运行为第 100 位。