当前位置:
X-MOL 学术
›
Comput. Oper. Res.
›
论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Two-phase algorithm for solving the preference-based multicriteria optimal path problem with reference points
Computers & Operations Research ( IF 4.6 ) Pub Date : 2020-09-01 , DOI: 10.1016/j.cor.2020.104977 Luigi Di Puglia Pugliese , Janusz Granat , Francesca Guerriero
Computers & Operations Research ( IF 4.6 ) Pub Date : 2020-09-01 , DOI: 10.1016/j.cor.2020.104977 Luigi Di Puglia Pugliese , Janusz Granat , Francesca Guerriero
Abstract The shortest path problem arises in several contexts like transportation, telecommunication or data analysis. New requirements in solving practical problems (e.g., efficient content delivery for information-centric networks, urban passenger transport system or social network) impose that more than one criterion should be considered. Since the objectives are in conflict, the solution is not unique, rather a set of (efficient) paths is defined as optimal. The most satisfactory path should be selected considering additional preference information. Generally, computing the entire set of efficient solutions is time consuming. In this paper, we apply the reference point method for finding the best path. In a reference point-based approach, non-additive scalarizing function is applied. In this case, the classical optimality principle for the shortest path problem is not valid. To overcome this issue, we propose an equivalent formulation dealing with the constrained shortest path (CSP) problem. The idea is to define a set of constraints guaranteeing that an optimal solution to the problem at hand lies in the feasible region of the defined CSP problem. We propose a two-phase method where, in the first phase, a bound on the optimal solution is computed and used to define the constraints, whereas, in the second phase a labelling algorithm is applied to search for an optimal solution to the defined CSP problem. The method is tested on instances generated from random and grid networks, considering several scenarios. The computational results show that, on average, the proposed solution strategy is competitive with the state-of-the-art approaches and behaves the best on grid networks.
中文翻译:
求解带参考点的基于偏好的多准则最优路径问题的两阶段算法
摘要 最短路径问题出现在多种环境中,如交通、电信或数据分析。解决实际问题的新要求(例如,以信息为中心的网络、城市客运系统或社交网络的高效内容交付)要求应考虑多个标准。由于目标存在冲突,因此解决方案不是唯一的,而是将一组(有效)路径定义为最佳。应该考虑额外的偏好信息来选择最令人满意的路径。通常,计算整套有效解决方案是耗时的。在本文中,我们应用参考点方法来寻找最佳路径。在基于参考点的方法中,应用了非加性标化函数。在这种情况下,最短路径问题的经典最优性原理是无效的。为了克服这个问题,我们提出了一种处理约束最短路径 (CSP) 问题的等效公式。这个想法是定义一组约束,保证手头问题的最佳解决方案位于定义的 CSP 问题的可行区域内。我们提出了一种两阶段方法,其中,在第一阶段,计算最优解的界限并用于定义约束,而在第二阶段,应用标记算法来搜索定义的 CSP 的最优解问题。该方法在随机和网格网络生成的实例上进行了测试,考虑了几种情况。计算结果表明,平均而言,
更新日期:2020-09-01
中文翻译:
求解带参考点的基于偏好的多准则最优路径问题的两阶段算法
摘要 最短路径问题出现在多种环境中,如交通、电信或数据分析。解决实际问题的新要求(例如,以信息为中心的网络、城市客运系统或社交网络的高效内容交付)要求应考虑多个标准。由于目标存在冲突,因此解决方案不是唯一的,而是将一组(有效)路径定义为最佳。应该考虑额外的偏好信息来选择最令人满意的路径。通常,计算整套有效解决方案是耗时的。在本文中,我们应用参考点方法来寻找最佳路径。在基于参考点的方法中,应用了非加性标化函数。在这种情况下,最短路径问题的经典最优性原理是无效的。为了克服这个问题,我们提出了一种处理约束最短路径 (CSP) 问题的等效公式。这个想法是定义一组约束,保证手头问题的最佳解决方案位于定义的 CSP 问题的可行区域内。我们提出了一种两阶段方法,其中,在第一阶段,计算最优解的界限并用于定义约束,而在第二阶段,应用标记算法来搜索定义的 CSP 的最优解问题。该方法在随机和网格网络生成的实例上进行了测试,考虑了几种情况。计算结果表明,平均而言,