Abstract:

The scheduling optimization of large-scale maintenance tasks has extensive applications in practical production processes such as optimizing maintenance scheduling for coal bed methane wells, well repair operations scheduling, and fracturing operations scheduling. This problem is large-scale and difficult to solve, which is a difficulty and challenge for real-time scheduling optimization. A well-designed schedule for large-scale maintenance tasks is of significant importance for ensuring smooth production and reducing costs in oil and gas fields. To effectively address this issue, an optimization algorithm based on ALNS-TS has been proposed, and its effectiveness has been verified through cases of different scales. The experimental results demonstrate that the solving time for representative cases with 10, 50 and 100 maintenance tasks is 0.03, 8.33, and 74.32 s, respectively, providing reasonable scheduling solutions within minutes. As the problem scale increases, the ALNS-TS based algorithm outperforms traditional algorithms in efficiency and is capable of finding lower objective function values and optimal solutions.

Key words: adaptive large neighborhood search, tabu search, maintenance scheduling, algorithm, optimization, systems engineering

摘要：

大规模维修任务的调度优化在实际生产过程中具有广泛的应用，例如煤层气井维修任务调度优化、修井作业调度和压裂作业调度等。该问题规模庞大且求解困难，是实时调度优化的难点和挑战。合理的大规模维修任务调度对于保障油气田平稳生产和降低成本具有重要意义。为了有效解决这一难题，提出了基于ALNS-TS的优化求解算法，并通过不同规模的案例验证了算法的有效性。实验结果显示，对于代表性的10、50和100个维修任务的案例，求解时间分别为0.03、8.33和74.32 s，都能在分钟级时间内给出合理的调度方案。随着问题规模增加，基于ALNS-TS的算法比传统算法更高效，并能找到目标函数值更低的更优解。

关键词: 自适应大邻域搜索, 禁忌搜索, 维修调度, 算法, 优化, 系统工程