New large-scale data instances for CARP and new variations of CARP

The capacitated arc routing problem (CARP) captures important aspects of real-life problems and has been studied extensively over the past two decades. Based on a waste collection project, we introduce a number of new CARP variations. We first present three multi-compartment CARP variations of different levels of complexity regarding compartments and where one incorporates a time horizon. We then present a variation that seeks to coordinate vehicles over a planning horizon such that the vehicles that collect different waste fractions from the same households do so on the same day of the week. Finally, the semi-periodic CARP takes into account that the households on a street, providing the demand of the edge, may not request waste collection at the same interval. We present large-scale instances both for the classical CARP and for the five new problems. The instances are based on real-life networks and waste data from five areas in Denmark and cover rural as well as urban areas. The largest instances contain more than 10,000 nodes. We give detailed information about the construction of the instances from the real-life data, and explain how they can be used to perform scenario analyses.

带容量弧路由问题（Capacitated Arc Routing Problem, CARP）涵盖了现实场景中的诸多核心问题特征，在过去二十年间得到了广泛深入的研究。本研究基于一项垃圾收集项目，提出了五类全新的CARP变体问题。首先，我们推出三类针对舱室配置具有不同复杂度的多舱室CARP变体，其中一类纳入了时间跨度约束；其次，我们提出一款协调调度型CARP变体：在规划周期内，从同一住户收集不同分类垃圾的清运车辆需于每周的同一天开展作业；最后，半周期CARP变体考虑到作为路段需求来源的沿街住户，其垃圾清运需求的间隔周期并不统一。我们构建了针对经典CARP以及上述五类新型问题的大规模测试算例集，该算例集基于丹麦五个区域的真实路网与垃圾清运数据，覆盖乡村与城市两类场景，其中规模最大的算例包含超过10000个节点。本文详细说明了基于真实数据构建测试算例的具体流程，并阐释了该算例集可用于开展场景化分析的应用方式。