Dream-Optimization-Algorithm-DOA-

DOA

As optimization problems grow increasingly complex, traditional deterministic algorithms often struggle to address these challenges. Metaheuristic algorithms, with their flexibility and low problem dependency, have emerged as a competitive alternative. This paper introduces the Dream Optimization Algorithm (DOA), inspired by human dreams, which exhibit partial memory retention, forgetting, and logical self-organization characteristics that bear strong similarities to the optimization process in metaheuristic algorithms. DOA incorporates a foundational memory strategy, a forgetting and supplementation strategy to balance exploration and exploitation, and a dream-sharing strategy to improve the ability to escape local optima. The optimization process is divided into exploration and exploitation phases, yielding satisfactory optimization results. This paper qualitatively analyzes DOA's search history, exploration--exploitation capabilities, and population diversity, showing its ability to adapt to problems of varying complexity. Quantitative analysis using three CEC benchmarks (CEC2017, CEC2019, CEC2022) compares DOA against 27 algorithms, including CEC2017 champion algorithms. Results indicate that DOA outperforms all competitors, showcasing superior convergence, advancement, stability, adaptability, robustness, significance, and reliability. Additionally, DOA achieved optimal results in eight engineering constrained optimization problems and in the practical application of photovoltaic cell model parameter optimization, demonstrating its effectiveness and practicality.