New random approaches of modified adaptive bats sonar algorithm for reservoir operation optimization problems / Nor Shuhada Binti Ibrahim

Faculty of Manufacturing & Mechatronics Engineering Technology

Thesis (Doctor of Philosophy) -- Universiti Malaysia Pahang – 2024

Includes bibliographical reference

The increasing interest among researchers in the application of metaheuristic algorithms for search optimization has resulted in notable progress, especially in tackling singleobjective optimization problems. However, the complexity significantly rises when dealing with multi objective functions. The Modified Adaptive Bats Sonar Algorithm (MABSA), initially designed for single objective optimization and inspired by colony bats' echolocation, has demonstrated efficiency with its simple structure and reduced computation time. This PhD thesis introduces an extended version of MABSA aimed at addressing constrained multi objective optimization problems by incorporating innovative random approaches, focusing to solves reservoir optimization problems. The research unfolds in four phases. Firstly, it delves into the development of reservoir operation optimization problems (ROOPs), which focusing on Malaysia-based model. Subsequently, it is exploring into the development of random approaches based on MABSA, integrating components into others algorithm whereby Squirrel Search Algorithm (SSA), Cauchy Mutation (CM) approach, and Doppler Effect (DE) phenomenon. Thirdly, the thesis validates the algorithm's performance on standard constrained single objective and multi objective benchmark test functions. In the fourth phase, the newly developed algorithm undergoes testing on the formulated ROOPs and compared to several contemporary optimizer algorithms. The significance of this research is underscored by the results obtained across all phases, which substantiate the newfound capability of the random approaches within MABSA to proficiently addressed into multi objective optimization problems. All the methods and data analysis were carried out using MATLAB simulator. Generally, the average results indicates that the newly developed algorithm retrieved better than original MABSA in finding a minimum point which increment 5.833% better than original MABSA, and new algorithm retrieved 5.204% more better than global optimum value. The significance of this research is underscored by the results obtained across all phases, which substantiate the newfound capability of the random approaches within MABSA to proficiently addresses especially in solving multi objective optimization problems. The algorithm emerges as highly competitive, consistently outperforming other existing metaheuristic algorithms in various scenarios, thus contributing significantly to the field of optimization research