Gao Yang, Cheng Liang
School of Petroleum Engineering, Yangtze University, Wuhan, China.
PLoS One. 2025 Jun 3;20(6):e0322111. doi: 10.1371/journal.pone.0322111. eCollection 2025.
Optimizing oil and gas production is of paramount importance in the petroleum sector, as it ensures the economic success of oil companies and meets the growing global demand for energy. The optimization of subsurface oil and gas production is critical for decision-makers, as it determines essential strategies like optimal well placement and well control parameters. Traditional reservoir production optimization methods often involve high computational costs and difficulties in achieving effective optimization. Evolutionary algorithms, inspired by biological evolution, have proven to be powerful tools for solving complex optimization challenges due to their independence from gradient information and efficient parallel processing capabilities. This paper proposes a highly efficient evolutionary algorithm for global optimization and oil and gas production optimization by enhancing the optimization performance of fruit fly optimization algorithm (FOA) through multi-swarm mechanism and greedy selection mechanism, which balance the algorithm's search and development capabilities. Specifically, after updating the population of FOA, we first apply multi-swarm mechanism to help the population escape local optima and improve the algorithm's search ability, and then apply greedy selection mechanism to enhance the population's development potential. To verify the optimization performance of MGFOA, we conducted comprehensive experimental validations at IEEE CEC 2017 and IEEE CEC 2022, including ablation studies, scalability experiments, search trace visualizations, and comparisons with other similar algorithms. Finally, MGFOA significantly outperformed other comparable algorithms in oil and gas production optimization.
优化油气生产在石油行业至关重要，因为它确保了石油公司的经济成功，并满足了全球日益增长的能源需求。地下油气生产的优化对决策者至关重要，因为它决定了诸如最佳井位和井控参数等关键策略。传统的油藏生产优化方法通常涉及高昂的计算成本，且难以实现有效的优化。受生物进化启发的进化算法，由于其不依赖梯度信息和高效的并行处理能力，已被证明是解决复杂优化挑战的强大工具。本文通过多群体机制和贪婪选择机制提高果蝇优化算法（FOA）的优化性能，提出了一种用于全局优化和油气生产优化的高效进化算法，该机制平衡了算法的搜索和开发能力。具体而言，在更新FOA的种群后，我们首先应用多群体机制帮助种群逃离局部最优并提高算法的搜索能力，然后应用贪婪选择机制增强种群的发展潜力。为了验证MGFOA的优化性能，我们在IEEE CEC 2017和IEEE CEC 2022上进行了全面的实验验证，包括消融研究、可扩展性实验、搜索轨迹可视化以及与其他类似算法的比较。最后，MGFOA在油气生产优化方面显著优于其他可比算法。
