Bridging Local and Global Knowledge: Cascaded Mixture-of-Experts Learning for Near-Shortest Path Routing

TL;DR

This paper introduces a cascaded mixture-of-experts model for near-shortest path routing that adaptively balances local and global information, significantly improving accuracy in sparse networks while maintaining computational efficiency.

Contribution

The paper proposes a novel Ca-MoE architecture with adaptive inference and online meta-learning, enhancing routing performance and generalization across diverse graph densities.

Findings

Up to 29.1% accuracy improvement in sparse networks.

Maintains performance within 1%-6% of the theoretical upper bound.

Efficiently balances local and global features through cascaded experts.

Abstract

While deep learning models that leverage local features have demonstrated significant potential for near-optimal routing in dense Euclidean graphs, they struggle to generalize well in sparse networks where topological irregularities require broader structural awareness. To address this limitation, we train a Cascaded Mixture of Experts (Ca-MoE) to solve the all-pairs near-shortest path (APNSP) routing problem. Our Ca-MoE is a modular two-tier architecture that supports the decision-making for forwarder selection with lower-tier experts relying on local features and upper-tier experts relying on global features. It performs adaptive inference wherein the upper-tier experts are triggered only when the lower-tier ones do not suffice to achieve adequate decision quality. Computational efficiency is thus achieved by escalating model capacity only when necessitated by topological complexity, and parameter redundancy is avoided. Furthermore, we incorporate an online meta-learning strategy that facilitates independent expert fine-tuning and utilizes a stability-focused update mechanism to prevent catastrophic forgetting as new graph environments are encountered. Experimental evaluations demonstrate that Ca-MoE routing improves accuracy by up to 29.1% in sparse networks compared to single-expert baselines and maintains performance within 1%-6% of the theoretical upper bound across diverse graph densities.