Structured expert judgement to understand the intrinsic vulnerability of traffic networks

Abstract

The concept of intrinsic vulnerability of a traffic network is defined for the first time in this paper. Intrinsic vulnerability is the susceptibility to incidents characterised by a probability of occurrence in space and time of difficult estimation, which can result in considerable reduction or loss of the system functionality.Given the nature of this type of vulnerability, its assessment might arise as a major problem. Therefore, this paper investigates the assessment of the intrinsic vulnerability of a traffic network through a set of quantifiable indicators, i.e., accessibility and reliability. Moreover, it is of interest to determine whether the selected indicators are sufficient to assess the intrinsic vulnerability or if there is any significant missing aspect to be considered. A new methodology based on structured elicitation of multivariate uncertainty from experts is presented to address these issues, allowing the estimation of the intrinsic vulnerability, and its probabilistic relationship with the indicators accessibility and reliability. Although applied to the case of the metric intrinsic vulnerability, the proposed methodology emerges as an effective tool to understand other traffic descriptors of difficult evaluation such as resilience.

Elicitation of Rank Correlations with Probabilities of Concordance: Method and Application to Building Management

“…Clearly, two groups of experts arose: whereas experts B and C obtained high d-calibration scores (≥0.85), experts A, D, and E obtained lower scores (dCal < 0.66). Interestingly, all experts were 'better' calibrated than in other studies implementing dependence calibration (e.g., [20,46]). This observation should be taken cautiously due to the relatively small number of seed variables used, as well as the fact that they are not related to the problem at hand [20,53].…”

“…Because Cooke's method was not designed for scoring dependence assessments, another performance-based method was investigated: the dependence calibration (or d-calibration [19]). The latter has already been applied to a handful of real-life problems [20,45,46]. For the Gaussian copula, this quantity measures the "distance" between two correlation matrices.…”

“…To better address this lack of data, so-called Gaussian copula-based Bayesian networks (GCBN) are adopted in this research. Their formulation, detailed in the next section, enables the involvement of field experts for the quantification of the model through structured expert judgments (SEJ), as demonstrated by past implementations of GCBNs (e.g., [18][19][20]).…”

“…Therefore, this paper focuses on the development of a method for the assessment of (conditional) rank correlations by field experts, while less attention is devoted to the elicitation of the one-dimensional marginal distributions. In contrast with existing research, which has delved into the use of statistical [19,21] and conditional fractile estimates [18,20,22] approaches, the relevance of a third type of probabilistic assessment is hereby studied: probabilities of concordance. Given the assumptions underlying GCBNs, unconditional rank correlations can be retrieved from concordance probabilities using a set of closed-form relations, which are defined in Section 2.2.1.…”

Attractiveness of Bike-Sharing Stations from a Multi-Modal Perspective: The Role of Objective and Subjective Features

“…Objective factors include those aspects that are based on facts, and therefore can be unequivocally measured after assuming some premises and uncertainties. They can be either directly assessed, for example, walking distance, or indirectly assessed through related indicators [43], such as the number of road accidents as a proxy for road safety. Subjective factors cannot be unequivocally measured, as they depend on the perceptions and experiences of the actors involved.…”