Optimal management of mixed hydraulic barriers in coastal aquifers using multi-objective Bayesian optimization

Mixed hydraulic barriers is an effective method to control seawater intrusion (SWI), particularly in regions that suffer from water shortages. However, determining the optimal well locations and rates for injection abstraction challenging due computational burden resulting huge number of calls high-fidelity hydrogeological simulation model. To alleviate this issue, we utilized a constrained multi-objective Bayesian optimization (BO) approach optimize minimize total cost, aquifer salinity, salt-wedge length, while satisfying regional abstractions with acceptable salinity levels. BO useful optimizing computationally expensive problems few iterations by using surrogate model acquisition function. Despite being efficient tool, use field coastal management has not been explored. The proposed framework was evaluated on unconfined subjected three scenarios considering different physical technical constraints benchmarked against widely used robust NSGA-II (Non-dominated Sorting Genetic Algorithm II) method. results proved effectiveness achieving optimum mixed design much fewer runs variable density 350 evaluations yielded comparable 4150 NSGA-II. solutions were spatially well-distributed along approximated Pareto front. For same evaluations, hypervolume obtained larger 30%. Based scenarios, average amount required ranged 1.5% 25% injection. significant impact SWI management, but abstracted provides alternative source water. A sensitivity analysis conducted problem illustrate its efficiency omitting one at time assessing impacts objective constraint functions.