Reliability-integrated multi-objective optimization of offshore oil spill mechanical recovery and oily wastewater management

Efficient offshore oil spill response requires minimizing response time, operational cost, and residual oil while accounting for operational uncertainty caused by equipment aging and failures. In this study, equipment reliability was modeled using Weibull-based failure probabilities linked to service age and integrated into multi-objective optimization. Two reliability-informed frameworks were developed: the three-objective model (R-OPA-III), which implicitly includes failure and maintenance constraints, and the four-objective model (R-OPA-IV), which explicitly optimizes reliability as an additional objective. Both were benchmarked against deterministic models (OPA-II and OPA-III) under a hypothetical large-scale oil spill in western Canada. Compared with OPA-II and OPA-III, R-OPA-III showed superior consistency, with averages of 437 h of total response time, CAD 1.99 M of total operational cost, and 224.4 m 3 of total weathered oil volume, the lowest variability (σ = 9.3 h, CAD 0.018 M, and 7.6 m 3 ), and the shortest normalized Euclidean distance to the ideal point (0.786). R-OPA-IV further extended the Pareto front (132 solutions, hypervolume = 0.850), providing broader trade-offs for decision-makers. This flexibility enables timely recovery and reduced environmental risk under uncertainty. Overall, R-OPA-III offers stable performance for general application, while R-OPA-IV supports adaptive planning. These reliability-informed models enhance the practicality and resilience of offshore oil spill response decision-making.