The year/Independent research

Paper 2601.07348

Controlled Self-Evolution for Algorithmic Code Optimization

Published
Jan 2026
Research lab
Independent
Citations
11
GitHub
131 stars

01 In brief

Summary

The paper introduces Controlled Self-Evolution (CSE), a framework to improve the exploration efficiency of self-evolution methods for algorithmic code optimization.

CSE addresses three bottlenecks: initialization bias, uncontrolled stochastic evolution, and insufficient experience utilization.

It comprises three components: Diversified Planning Initialization, which generates structurally distinct algorithmic strategies; Genetic Evolution, which uses feedback-guided controlled mutation and compositional crossover; and Hierarchical Evolution Memory, which captures intra-task and inter-task experiences.

Experiments on EffiBench-X show CSE consistently outperforms baselines (Direct, Self-Reflection, SE-Agent, AlphaEvolve) across LLM backbones (Qwen3-235B-A22B, DeepSeek-v3-0324, Claude-4.5-Sonnet, GPT-5) in Python and C++, with higher efficiency from early generations and sustained improvement.

Ablations confirm all components are essential, with memory providing the largest impact.

The code is available at https://github.com/QuantaAlpha/EvoControl.

02 From the paper

Abstract

Self-evolution methods enhance code generation through iterative "generate-verify-refine" cycles, yet existing approaches suffer from low exploration efficiency, failing to discover solutions with superior complexity within limited budgets. This inefficiency stems from initialization bias trapping evolution in poor solution regions, uncontrolled stochastic operations lacking feedback guidance, and insufficient experience utilization across tasks. To address these bottlenecks, we propose Controlled Self-Evolution (CSE), which consists of three key components. Diversified Planning Initialization generates structurally distinct algorithmic strategies for broad solution space coverage. Genetic Evolution replaces stochastic operations with feedback-guided mechanisms, enabling targeted mutation and compositional crossover. Hierarchical Evolution Memory captures both successful and failed experiences at inter-task and intra-task levels. Experiments on EffiBench-X demonstrate that CSE consistently outperforms all baselines across various LLM backbones. Furthermore, CSE achieves higher efficiency from early generations and maintains continuous improvement throughout evolution. Our code is publicly available at https://github.com/QuantaAlpha/EvoControl.