Adaptive Conditional Forest Sampling for Spectral Risk Optimisation under Decision-Dependent Uncertainty
arXiv:2603.12507v1 Announce Type: new Abstract: Minimising a spectral risk objective, defined as a convex combination of expected cost and Conditional Value-at-Risk (CVaR), is challenging when the uncertainty distribution is decision-dependent, making both surrogate modelling and simulation-based ranking sensitive to tail estimation error. We propose Adaptive Conditional Forest Sampling (ACFS), a four-phase simulation-optimisation framework that integrates Generalised Random Forests for decision-conditional distribution approximation, CEM-guided global exploration, rank-weighted focused augmentation, and surrogate-to-oracle two-stage reranking before multi-start gradient-based refinement. We evaluate ACFS on two structurally distinct data-generating processes: a decision-dependent Student-t copula and a Gaussian copula with log-normal marginals, across three penalty-weight configurations and 100 replications per setting. ACFS achieves the lowest median oracle spectral risk on the seco
arXiv:2603.12507v1 Announce Type: new Abstract: Minimising a spectral risk objective, defined as a convex combination of expected cost and Conditional Value-at-Risk (CVaR), is challenging when the uncertainty distribution is decision-dependent, making both surrogate modelling and simulation-based ranking sensitive to tail estimation error. We propose Adaptive Conditional Forest Sampling (ACFS), a four-phase simulation-optimisation framework that integrates Generalised Random Forests for decision-conditional distribution approximation, CEM-guided global exploration, rank-weighted focused augmentation, and surrogate-to-oracle two-stage reranking before multi-start gradient-based refinement. We evaluate ACFS on two structurally distinct data-generating processes: a decision-dependent Student-t copula and a Gaussian copula with log-normal marginals, across three penalty-weight configurations and 100 replications per setting. ACFS achieves the lowest median oracle spectral risk on the second benchmark in every configuration, with median gaps over GP-BO ranging from 6.0% to 20.0%. On the first benchmark, ACFS and GP-BO are statistically indistinguishable in median objective, but ACFS reduces cross-replication dispersion by approximately 1.8 to 1.9 times on the first benchmark and 1.7 to 2.0 times on the second, indicating materially improved run-to-run reliability. ACFS also outperforms CEM-SO, SGD-CVaR, and KDE-SO in nearly all settings, while ablation and sensitivity analyses support the contribution and robustness of the proposed design.
Executive Summary
This study proposes Adaptive Conditional Forest Sampling (ACFS), a four-phase simulation-optimisation framework that integrates Generalised Random Forests for decision-conditional distribution approximation, CEM-guided global exploration, rank-weighted focused augmentation, and surrogate-to-oracle two-stage reranking before multi-start gradient-based refinement. ACFS achieves the lowest median oracle spectral risk on one benchmark and outperforms existing methods in nearly all settings across various configurations. However, the study's conclusions are limited to the specific data-generating processes and penalty-weight configurations explored. Further research is needed to evaluate ACFS in a broader range of scenarios and assess its applicability to real-world problems. The study offers a promising solution for spectral risk optimisation under decision-dependent uncertainty, but its practical and policy implications require further investigation.
Key Points
- ▸ ACFS integrates multiple methods to tackle decision-dependent uncertainty, including Generalised Random Forests and CEM-guided global exploration.
- ▸ ACFS outperforms existing methods in nearly all settings across various configurations.
- ▸ The study's conclusions are limited to the specific data-generating processes and penalty-weight configurations explored.
Merits
Strength in decision-dependent uncertainty handling
ACFS's use of Generalised Random Forests and CEM-guided global exploration enables effective handling of decision-dependent uncertainty.
Demerits
Limited generalisability
The study's conclusions are limited to the specific data-generating processes and penalty-weight configurations explored, which may not be representative of real-world scenarios.
Expert Commentary
The study's use of ACFS to tackle decision-dependent uncertainty is a significant contribution to the field of spectral risk optimisation. However, the study's conclusions are limited to the specific data-generating processes and penalty-weight configurations explored. Further research is needed to evaluate ACFS in a broader range of scenarios and assess its applicability to real-world problems. Additionally, the study's findings highlight the importance of considering decision-dependent uncertainty in spectral risk optimisation, which can inform policy decisions in areas such as risk management and regulatory compliance.
Recommendations
- ✓ Future research should investigate the generalisability of ACFS across various data-generating processes and penalty-weight configurations.
- ✓ The study's findings should be applied in real-world scenarios to assess the practicality and policy implications of ACFS.
