Modelling the reliability of search and rescue operations with Bayesian Belief Networks
Introduction
The UK government is concerned with improving the operational performance of the Maritime and Coastguard Agency (MCA). Since 1994, there have been a number of reviews of the effectiveness of Search And Rescue (SAR) operations coordinated by the MCA, culminating in the decision in 1999 to close a number of the Maritime Rescue coordination centres. However, there appears to have been a distinct lack of formal analysis to justify this decision.
The aim of this research is to conduct a formal analysis of the Coastguard SAR service in order to identify the key factors that influence its effectiveness and to provide a way of measuring this influence. An earlier study had been conducted with this aim, which made use of publicly available secondary data (in the form of annual Incident Statistics) and developed a binary logistic regression model to support the assessment [1]. The findings of this previous study highlighted the importance of explanatory variables such as the size or scale of incidents, staff workload and the length of coastline monitored by each coordination centre. Such variables may be interpreted as providing rather crude composite indicators for the multitude of factors that determine the effectiveness of this kind of complex operational system. The contribution of the present study is to support a more detailed explanation of the relationship between the reliability of SAR operations and identifiable causal factors.
In Section 2, we present a summary of background research. First, we present a summary of recent events relating to the UK Maritime and Coastguard Agency that provided the motivation for this research; and second, we present a summary of relevant research literature on maritime risk. In Section 3, we present a summary of our analysis of the available secondary data. This analysis is conducted at a crude aggregate level, but highlights some significant characteristics associated with better-performing centres. In Section 4, we describe the Bayesian Belief Network (BBN) that we obtained and reflect on the process of eliciting subjective probabilities. In Section 5, we conclude by reflecting on the synergies between statistical modelling and BBNs, whereby the former supports the validation of the latter and the latter provides a causal description of the associations identified in the former.
Section snippets
Background
Within this section, we provide a brief overview of the UK Maritime and Coastguard Agency as related to this research project; and in Section 2.2 we provide a brief summary of research conducted on maritime risk issues.
Secondary data analysis
This section will present a summary of the analysis conducted on the available secondary data. The original study by Van der Meer et al. [1] was based on data for the period 1995–1999; the present study uses data for the period 1995–2004. We explore the data through contingency tables, which can then form the basis for an empirical BBN. In Section 3.2, we shall provide a discussion of the shortcomings of the data and the implications for inference.
Elicitation of BBN
Within this section BBNs are briefly described, followed by the description of the elicitation process undertaken to develop the BBN for this situation. The final model is shown along with a reflection on the process of elicitation.
Summary and future work
The BBN analysis concerning environmental factors identified several other factors that the UK Coastguard do not officially measure. However, captured within the BBN are variables related to staff workload, local knowledge and size and severity of incidents that are consistent with the statistical analysis described in Section 3.
The process of constructing the BBN resulted in a deeper understanding of validity (and lack thereof) of the surrogate measures used with the secondary data analysis.
References (27)
Determination of probability of marine accidents with respect to gas carriers proceeding in Dutch Coastal and Inland Waters
J Hazard Mater
(1980)- et al.
GRACAT: software for grounding and collision risk analysis
Mar Struct
(2002) Collision scenarios and probabilistic collision damage
Mar Struct
(2002)An interactive multiobjective model for the strategic maritime transportation of petroleum products: risk analysis and routing
Saf Sci
(2001)- et al.
POLSSS: policy making for sea shipping safety
Saf Sci
(2000) POLSSS: surveying stakeholders about acceptability of risks and system changes
Saf Sci
(2000)POLSSS: overview and cost-effectiveness analysis
Saf Sci
(2000)- et al.
Study of a collision-avoidance system for ships
Control Eng Pract
(1998) - et al.
Risk based methodology for safety improvements in ports
J Hazard Mater
(2000) - et al.
A comprehensive assessment system for the maritime traffic environment
Saf Sci
(1995)
Risk analysis in harbor environments using simulation
Saf Sci
The current status and future aspects in formal ship safety assessment
Saf Sci
Application of Bayesian network to the probabilistic risk assessment of nuclear waste disposal
Reliab Eng Syst Saf
Cited by (82)
Exploring the relative importance of sustainable competitiveness pillars
2024, Journal of Cleaner ProductionEmergency response in cascading scenarios triggered by natural events
2024, Reliability Engineering and System SafetyData-driven Bayesian network approach to maritime accidents involved by dry bulk carriers in Turkish search and rescue areas
2023, Regional Studies in Marine ScienceRisk assessment of the Ship steering gear failures using fuzzy-Bayesian networks
2023, Ocean EngineeringOptimization of maritime support network with relays under uncertainty: A novel matheuristics method
2023, Reliability Engineering and System Safety