Samare Hashemi Syed Mohsen, Robati Amir, Kazerooni Mohammad Ali
Department of Civil Engineering, Islamic Azad University-Kerman Branch, Kerman, Iran.
Sci Rep. 2024 Feb 13;14(1):3607. doi: 10.1038/s41598-024-54326-z.
The optimal operation of the multi-purpose reservoir system is a difficult, and, sometimes, non-linear problem in multi-objective optimization. By simulating biological behavior, meta-heuristic algorithms scan the decision space and can offer a set of points as a group of solutions to a problem. Because it is essential to simultaneously optimize several competing objectives and consider relevant constraints as the main problem in many optimization problems, researchers have improved their ability to solve multi-objective problems by developing complementary multi-objective algorithms. Because the AHA algorithm is new, its multi-objective version, MOAHA (multi-objective artificial hummingbird algorithm), was used in this study and compared with two novel multi-objective algorithms, MOMSA and MOMGA. Schaffer and MMF1 were used as two standard multi-objective benchmark functions to gauge the effectiveness of the proposed method. Then, for 180 months, the best way to operate the reservoir system of the Karun River basin, which includes Karun 4, Karun 3, Karun 1, Masjed-e-Soleyman, and Gotvand Olia dams to generate hydropower energy, supply downstream demands (drinking, agriculture, industry, environmental), and control flooding was examined from September 2000 to August 2015. Four performance appraisal criteria (GD, S, Δ, and MS) and four evaluation indices (reliability, resiliency, vulnerability, and sustainability) were used in Karun's multi-objective multi-reservoir problem to evaluate the performance of the multi-objective algorithm. All three algorithms demonstrated strong capability in criterion problems by using multi-objective algorithms' criteria and performance indicators. The large-scale (1800 dimensions) of the multi-objective operation of the Karun Basin reservoir system was another problem. With a minimum of 1441.71 objectives and an average annual hydropower energy manufacturing of 17,166.47 GW, the MOAHA algorithm demonstrated considerable ability compared to the other two. The final results demonstrated the MOAHA algorithm's excellent performance, particularly in difficult and significant problems such as multi-reservoir systems' optimal operation under various objectives.
多功能水库系统的优化运行是多目标优化中一个困难且有时是非线性的问题。通过模拟生物行为，元启发式算法扫描决策空间，并能提供一组点作为问题的一组解决方案。由于在许多优化问题中，同时优化几个相互竞争的目标并考虑相关约束是主要问题，研究人员通过开发互补的多目标算法提高了他们解决多目标问题的能力。由于AHA算法是新算法，其多目标版本MOAHA（多目标人工蜂鸟算法）在本研究中被使用，并与两种新颖的多目标算法MOMSA和MOMGA进行比较。Schaffer和MMF1被用作两个标准多目标基准函数来衡量所提方法的有效性。然后，在2000年9月至2015年8月期间，研究了卡伦河流域水库系统（包括卡伦4号、卡伦3号、卡伦1号、马斯吉德苏莱曼和戈特万德奥利亚大坝）的最佳运行方式，以生产水电能源、满足下游需求（饮用水、农业、工业、环境）并控制洪水。在卡伦的多目标多水库问题中，使用了四个性能评估标准（GD、S、Δ和MS）和四个评估指标（可靠性、恢复力、脆弱性和可持续性）来评估多目标算法的性能。通过使用多目标算法的标准和性能指标，所有三种算法在标准问题上都表现出强大的能力。卡伦河流域水库系统多目标运行的大规模（1800维）是另一个问题。MOAHA算法在至少1441.71个目标且年均水电能源产量为17166.47吉瓦的情况下，与其他两种算法相比表现出相当强的能力。最终结果表明MOAHA算法具有优异的性能，特别是在诸如多水库系统在各种目标下的优化运行等困难且重要的问题中。
