An Original Contribution to Bayesian Decision Analysis Theory for Field Development Optimisation Under Geological Uncertainty in Nigerian Deepwater Frontier Blocks

Field development decision-making under geological uncertainty in frontier deepwater blocks involves high-stakes choices between development concept options, well count, and phasing strategies where the value of information and the optimal decision sequence cannot be determined without a formal probabilistic decision framework. This study makes an original contribution to Bayesian decision analysis theory applied to field development optimisation in Nigerian deepwater frontier blocks. The theoretical contribution develops a multi-stage influence diagram formulation that explicitly represents geological uncertainty through a hierarchically structured Bayesian network linking seismic attribute observations, depositional system model parameters, and reservoir performance outcomes. An original algorithm for backward induction through the multi-stage influence diagram is derived, incorporating sampling-based rollback for problems that are analytically intractable due to continuous observation spaces. The framework is applied to a Nigerian deepwater frontier block development decision problem with three reservoir appraisal stages and a final development concept selection. Decision tree pruning algorithms and importance sampling reduce the computational burden from 2.4 million scenario evaluations required for full enumeration to 48,000 scenarios without loss of expected value accuracy exceeding 2 percent. Sensitivity analysis identifies that the optimal expected value of perfect information for the primary development concept selection is 180 million USD, while targeted value of information for a single appraisal well ranges from 40 to 95 million USD depending on well location within the structural closure. Keywords: Bayesian decision analysis, field development optimisation, deepwater frontier blocks, geological uncertainty, value of information.