Problem
Research questions and friction points this paper is trying to address.
Black-box Optimization
Public Policy
Complex Problem Solving
Black-Box Optimization with Implicit Constraints for Public Policy
🤖 AI Summary
Black-box optimization (BBO) in public policy—exemplified by precinct boundary delineation—faces challenges including high-dimensional decision spaces, implicit constraints that are difficult to model and search over, and the absence of gradient information.
Method: This paper proposes CageBO, a novel BBO framework that introduces conditional variational autoencoders (CVAEs) for the first time in this context. CageBO learns a distribution over feasible decisions and establishes a bijective, invertible mapping between the constrained original space and an unconstrained latent space, enabling efficient black-box optimization exclusively in the latent space. Crucially, it implicitly captures domain-specific rules without requiring explicit constraint modeling.
Results: Evaluated on a large-scale, real-world precinct delineation task in Atlanta, CageBO substantially outperforms state-of-the-art baselines, achieving superior trade-offs between solution quality and computational efficiency. It establishes a scalable, interpretable BBO paradigm tailored for policy-driven complex decision-making.
📝 Abstract
Black-box optimization (BBO) has become increasingly relevant for tackling complex decision-making problems, especially in public policy domains such as police redistricting. However, its broader application in public policymaking is hindered by the complexity of defining feasible regions and the high-dimensionality of decisions. This paper introduces a novel BBO framework, termed as the Conditional And Generative Black-box Optimization (CageBO). This approach leverages a conditional variational autoencoder to learn the distribution of feasible decisions, enabling a two-way mapping between the original decision space and a simplified, constraint-free latent space. The CageBO efficiently handles the implicit constraints often found in public policy applications, allowing for optimization in the latent space while evaluating objectives in the original space. We validate our method through a case study on large-scale police redistricting problems in Atlanta, Georgia. Our results reveal that our CageBO offers notable improvements in performance and efficiency compared to the baselines.
Innovation
Methods, ideas, or system contributions that make the work stand out.
CageBO
Policy-making Optimization
Conditional Generative Black-box Optimization
🔎 Similar Papers
No similar papers found.
