TOPAS Scheduler Balances Prefix Caching and Batching for Multi-Agent LLMs

TOPAS is a workflow-aware scheduler designed to optimize multi-agent LLM serving under shared KV cache constraints. It optimizes task-level job completion time by jointly deciding which agent prefixes to retain in cache and which requests to schedule for execution.