A Markov Chain Approach to Multicriteria Decision Analysis with an Application to Offshore Decommissioning

Abstract

This paper proposes a novel approach that makes use of continuous-time Markov chains and regret functions to find an appropriate compromise in the context of multicriteria decision analysis (MCDA). This method was an innovation in the relationship between uncertainty and decision parameters, and it allows for a much more robust sensitivity analysis. The proposed approach avoids the drawbacks of arbitrary user-defined and method-specific parameters by defining transition rates that depend only upon the performances of the alternatives. This results in a flexible and easy-to-use tool that is completely transparent, reproducible, and easy to interpret. Furthermore, because it is based on Markov chains, the model allows for a seamless and innovative treatment of uncertainty. We apply the approach to an oil and gas decommissioning problem, which seeks a responsible manner in which to dismantle and deactivate production facilities. The experiments, which make use of published data on the decommissioning of the field of Brent, account for 12 criteria and illustrate the application of the proposed approach.