Scenario-based optimization problems can be solved via Benders decomposition, which separates first-stage (master problem) decisions from second-stage (subproblem) recourse actions and iteratively refines the master problem with Benders cuts. In conventional Benders decomposition, all subproblems are solved at each iteration. For problems with many scenarios, solving only a selected subset can reduce computation. We quantify the potential in selecting only those subproblems that yield cuts, and develop subproblem scoring and selection strategies. The proposed multi-criteria scoring methods combine historical subproblem performance metrics with problem-specific features, trained online via logistic regression to adapt to the changing likelihood of subproblem usefulness. Multiple stopping criteria balance exploration and exploitation: cut limits, proportional solve limits, and score thresholds. We evaluate our approach on a variant of the survivable network design problem, which serves as a testbed due to its natural decomposition into many subproblems of varying importance.
Computational experiments on 135 test instances demonstrate the potential and practical performance of subproblem selection. Analysis reveals that 52.1% of all subproblems solved are unnecessary (they contribute no cuts and occur outside cut-free rounds). An oracle with perfect foresight reduces total solve times by 34.4%. Random selection performs significantly worse than full enumeration, showing that naive strategies can degrade performance. Our best-scoring and selection method achieves statistically significant improvements in both runtime and primal-dual integrals. These results provide empirical evidence that informed subproblem selection can improve Benders decomposition in this setting, while highlighting challenges in developing reliable prediction models. Whether these findings extend to other problem classes remains an open question for future work.