Maritime ports are an integral part of global trade and the supply network system. An upcoming paradigm for innovation in this system is that of the Physical Internet (PI). This highly advanced way of shipping will present a very different logistics environment with respective challenges for maritime ports. For those investing in or operating port systems, it is important to understand whether different service quality aspects will be important in this future system, compared to today. Our paper deals with the port selection problem. Although it has been studied extensively in a contemporary context, there has been no exploration of the criteria and preferences of decision-makers in the future shipping environment of the PI. Our objective is to define these criteria and explore their weighting in this new context. We propose two distinct autonomous decision-makers for port performance evaluation and selection in the PI: intelligent containers and vessels. We identify future port performance evaluation and selection criteria, and analyse their weighting based on an expert survey, complementing the extant literature on port selection and the PI. We use the Bayesian Best-Worst Method (BWM) to derive weights for the criteria. We find that, compared to the current port performance evaluation and selection literature, in a first stage in the modelling of intelligent agents’ performance preferences, subtle differences in weights mark the step from the present towards the PI. Partly, this is reassuring for port authorities as they can manage largely the same set of performance indicators to be attractive for both decision-makers. However, the results also show differences between agents, with an increased importance of, in particular, Level of Service, Network Interconnectivity, and Information Systems.

海运港口是全球贸易和供应网络系统的组成部分。该系统即将到来的创新范式是物理互联网 (PI)。这种高度先进的航运方式将为海港带来截然不同的物流环境和各自的挑战。对于那些投资或运营港口系统的人来说，重要的是要了解与今天相比，不同的服务质量方面在这个未来系统中是否重要。我们的论文涉及端口选择问题。尽管在当代背景下对其进行了广泛的研究，但尚未对 PI 未来航运环境中决策者的标准和偏好进行探索。我们的目标是定义这些标准并探索它们在这种新环境中的权重。我们为 PI 中的港口绩效评估和选择提出了两个不同的自主决策者：智能集装箱和船舶。我们确定未来港口绩效评估和选择标准，并根据专家调查分析它们的权重，补充有关港口选择和 PI 的现有文献。我们使用贝叶斯最佳-最差方法 (BWM) 来推导标准的权重。我们发现，与当前的端口性能评估和选择文献相比，在智能代理性能偏好建模的第一阶段，权重的细微差异标志着从现在到 PI 的步骤。在某种程度上，这让港口当局放心，因为他们可以管理大致相同的一组绩效指标，从而对双方决策者都具有吸引力。然而，