The manufacturing of oil extraction pipes involves several steps of pipe processing in a sequence of machines. The pipes are attached to reels that are moved between the machines by cranes. As the physical space and the reach of the cranes are limited, the motion must follow specific paths, which must comply with several movement constraints. A poorly designed reel movement plan will require a long time for reel positioning and can even lead the facility to a deadlock. This work proposes a reel management heuristic that, given the current reel positions, the facility configuration, and the production plan to be executed, builds a full reel movement plan in a few minutes. The proposed tool optimizes five objectives that are combined into a lexicographical function: task compliance; moves to the uncovered part of the facility; tardiness; earliness; and the number of reel movements performed. The optimized movements are generated by a fast local search based heuristic, which employs a distance metric suitable for measuring distances in this study. That distance metric, which is inspired on the Dijkstra’s algorithm, measures the effort for moving a reel from a start vertex to a target vertex. The proposed heuristic for movement planning is conceived as a real-time tool, which is re-run on each time a relevant discrepancy between the planning and the actual behavior of the plant appears; for this reason, the proposed system is allowed a maximum of 300 s of processing time for delivering a complete solution. Comparisons with a heuristic based on ILP and with a VNS procedure were performed, showing that the proposed method reaches either the same performance or a better performance in the objectives, consuming a much smaller processing time.

石油开采管的制造涉及在一系列机器中进行管材加工的几个步骤。管道连接到由起重机在机器之间移动的卷轴。由于起重机的物理空间和范围是有限的，运动必须遵循特定的路径，这必须符合几个运动约束。设计不当的卷轴移动计划将需要很长时间来定位卷轴，甚至可能导致设施陷入僵局。这项工作提出了一种卷轴管理启发式方法，根据当前卷轴位置、设施配置和要执行的生产计划，在几分钟内构建完整的卷轴移动计划。建议的工具优化了五个目标，这些目标组合成一个词典功能：任务合规性；移动到设施未覆盖的部分；迟到；早日; 以及执行的卷轴运动次数。优化的运动由基于快速局部搜索的启发式生成，该启发式采用适合在本研究中测量距离的距离度量。该距离度量受到 Dijkstra 算法的启发，测量将卷轴从起始顶点移动到目标顶点的努力。提议的运动规划启发式被认为是一种实时工具，每次出现规划与工厂实际行为之间的相关差异时都会重新运行；出于这个原因，建议的系统最多允许 300 秒的处理时间来提供完整的解决方案。与基于 ILP 的启发式算法和 VNS 程序进行了比较，