Marine transportation application of risk-based decision making

Decision making frequently involves determining parameters that are imbued with uncertainty. This sometimes involves the use of subjective, qualitative methods, such as a risk ranking matrix, in which consideration is given both to the probability of occurrence and the expected consequences of an event. These are found on the horizontal and vertical axes of the matrix; inside the matrix are blocks denoting the expected severity of the situation, with each block denoting risk levels, such as acceptable, moderate, serious, and critical. The decision maker makes a qualitative determination of both input parameters, and where the two intersect on the matrix determines the severity of the situation, which informs the action to be taken. The inputs (probability of loss and resulting damage) are typically imprecise. Recent research has criticized methods such as the risk ranking matrix as being ineffective and for giving users a false sense of security, which can have serious consequences at sea.

One alternative is the use of Monte Carlo (MC) simulations. MC simulation is a quantitative risk analysis technique where the inputs, such as the probability of a loss event, are modelled as statistical probability density functions (PDFs) rather than given imprecise labels, such as low, medium, or high. Once the inputs have been characterized as PDFs, a MC simulation program can determine the expected outcome thousands of times using random numbers along with the PDFs to determine the actual values of each input parameter for each particular instance the simulation is run. This will result in thousands of outcomes being generated. The aggregation of these outcomes allows the decision maker to determine the outcomes for the worst case scenario, the best case scenario, and the most likely scenario, along with the statistical probability of each scenario. Such a tool is more powerful and informative than a risk ranking matrix.

The paper begins with an overview of risk ranking matrices and associated problems. Next, we provide an overview of Monte Carlo simulation and explain its use in marine risk management situations. We then present a hypothetical case in which a Monte Carlo simulation is used to advise the course of action for a shipping company considering using the Northern Sea Route instead of the Suez Canal for shipping between Rotterdam and Yokohama. We conclude that the use of Monte Carlo simulation is a promising option for risk-based decision making at sea, that significant work is required in the area of characterizing input parameters as PDFs, and that training in the areas of probability and statistics should be an important part of the curriculum at MET institutions.