Bregman Centroid Guided Cross-Entropy Method

The Cross-Entropy Method (CEM) is a widely adopted trajectory optimizer in model-based reinforcement learning (MBRL), but its unimodal sampling strategy often leads to premature convergence in multimodal landscapes. In this work, we propose Bregman Centroid Guided CEM ($\mathcal{B}\mathcal{C}$-EvoCEM), a lightweight enhancement to ensemble CEM that leverages $\mathit{\text{Bregman}}\mathit{\text{centroids}}$ for principled information aggregation and diversity control. \textbf{$\mathcal{B}\mathcal{C}$-EvoCEM} computes a performance-weighted Bregman centroid across CEM workers and updates the least contributing ones by sampling within a trust region around the centroid. Leveraging the duality between Bregman divergences and exponential family distributions, we show that \textbf{$\mathcal{B}\mathcal{C}$-EvoCEM} integrates seamlessly into standard CEM pipelines with negligible overhead. Empirical results on synthetic benchmarks, a cluttered navigation task, and full MBRL pipelines demonstrate that \textbf{$\mathcal{B}\mathcal{C}$-EvoCEM} enhances both convergence and solution quality, providing a simple yet effective upgrade for CEM.