Router DC

Overview

router_dc is a selection algorithm for semantic query-to-model matching.

It aligns to config/algorithm/selection/router-dc.yaml.

Key Advantages

• Uses semantic similarity instead of only explicit ranking rules.
• Keeps learned selector thresholds visible in config.
• Useful when prompt semantics matter more than static priority.

What Problem Does It Solve?

Some routes need the chosen model to match the prompt style or task semantics closely. router_dc makes that learned semantic selector part of the decision instead of an external hidden step.

When to Use

• the best candidate depends on semantic similarity between prompt and model profile
• you want a learned selector without full online exploration
• one route should route by semantic fit rather than only cost or latency

Configuration

Use this fragment inside routing.decisions[].algorithm:

algorithm:
  type: router_dc
  router_dc:
    temperature: 0.2
    dimension_size: 384
    min_similarity: 0.7
    use_query_contrastive: true