Improving Coherence and Persistence in Agentic AI for System Optimization

2026-03-22

AI Summary

Existing agents are constrained by limited context windows, hindering their ability to maintain coherent long-term exploration and accumulate knowledge in complex system optimization, often leading to suboptimal local solutions. This work proposes the Engram architecture, which decouples the exploration process into multiple iterative rounds of agent design, testing, and analysis, and introduces a persistent Research Digest mechanism to enable cross-iteration inheritance of high-level knowledge. By transcending the limitations of single-context interactions, this approach significantly enhances agent coherence and continual learning capabilities in multi-step collaborative optimization. Empirical results demonstrate substantial performance improvements across diverse tasks, including multi-cloud multicast scheduling, LLM inference request routing, and KV cache reuse under natural language queries.

Technology Category

• Cognitive Modeling & Cognitive Systems: Agent Architectures
• Multiagent Systems: Mechanism Design
• Search and Optimization: Learning to Search

Application Category

• Search and Retrieval-Augmented AI: Agentic search
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Economics, Online Markets and Human Computation: Incentives in network design for Web infrastructures and ecosystems

Abstract

Designing high-performance system heuristics is a creative, iterative process requiring experts to form hypotheses and execute multi-step conceptual shifts. While Large Language Models (LLMs) show promise in automating this loop, they struggle with complex system problems due to two critical failure modes: evolutionary neighborhood bias and the coherence ceiling. Evolutionary methods often remain trapped in local optima by relying on scalar benchmark scores, failing when coordinated multi-step changes are required. Conversely, existing agentic frameworks suffer from context degradation over long horizons or fail to accumulate knowledge across independent runs. We present Engram, an agentic researcher architecture that addresses these limitations by decoupling long-horizon exploration from the constraints of a single context window. Engram organizes exploration into a sequence of agents that iteratively design, test, and analyze mechanisms. At the conclusion of each run, an agent stores code snapshots, logs, and results in a persistent Archive and distills high-level modeling insights into a compact, persistent Research Digest. Subsequent agents then begin with a fresh context window, reading the Research Digest to build on prior discoveries. We find that Engram exhibits superior performance across diverse domains including multi-cloud multicast, LLM inference request routing, and optimizing KV cache reuse in databases with natural language queries.

Problem

• Agentic AI
• Coherence
• Persistence
• System Optimization
• Large Language Models

Innovation

• agentic AI
• persistent memory
• research digest
• system optimization
• long-horizon reasoning