Kyriklidis Christos, Dounias Georgios
Department of Chemical Engineering, University of Western Macedonia, Kila, GR-50100, Kozani, Greece.
Management and Decision Engineering Laboratory, Department of Financial and Management Engineering, University of the Aegean, 41 Kountouriotou Str. GR-82100, Chios, Greece.
Data Brief. 2023 Jun 24;49:109340. doi: 10.1016/j.dib.2023.109340. eCollection 2023 Aug.
Resource leveling is a highly complex optimization problem corresponding to adjusting a project's timeline (start and end dates) with the aim of matching resource allocation demands. The problem is particularly complex when a project is large and involves hundreds or even thousands of activities. Its successful solution is equivalent to considerable profits for the involved construction groups through the efficient management of their resources. In literature usually can be found only small-size benchmark problems consisting of a few activities (i.e., ten to twenty) mainly aiming to demonstrate that a new proposed method can operate correctly identifying the optimum (or a near-optimum) solution. This data article provides resource leveling data suitable for testing, corresponding to a very large real-world problem of ship construction (consisting of 1178 activities). According to recent literature, the majority of the proposed methods for solving resource-leveling optimization problems are based on algorithmic approaches, usually artificial intelligence-oriented (evolutionary programming). The reason is that intelligent approaches manage to solve complex problems, producing approximate solutions of high accuracy and thus attractive (profitable) for practical application. The provided data have been tested in the past with intelligent techniques using different evaluation functions. Nevertheless, the specific dataset has never been published before elsewhere and now there is a clear opportunity to provide these data for testing and benchmark experimentation to interested researchers.
资源平衡是一个高度复杂的优化问题，它对应于调整项目的时间表（开始和结束日期），目的是匹配资源分配需求。当项目规模很大且涉及数百甚至数千个活动时，这个问题就特别复杂。其成功解决等同于通过有效管理资源，为相关建设集团带来可观的利润。在文献中通常只能找到由少数活动（即十到二十个）组成的小规模基准问题，主要目的是证明新提出的方法能够正确运行，识别出最优（或近似最优）解。本文提供了适合测试的资源平衡数据，对应一个非常大的船舶建造实际问题（由1178个活动组成）。根据最近的文献，大多数用于解决资源平衡优化问题的方法都基于算法方法，通常是面向人工智能的（进化规划）。原因是智能方法能够解决复杂问题，产生高精度的近似解，因此在实际应用中很有吸引力（有利可图）。所提供的数据过去曾使用不同的评估函数通过智能技术进行测试。然而，这个特定的数据集以前从未在其他地方发表过，现在有一个明确的机会为感兴趣的研究人员提供这些数据用于测试和基准实验。
