Technology Category
Search and Optimization: Metareasoning and MetaheuristicsConstraint Satisfaction and Optimization: Mixed Discrete/Continuous OptimizationReasoning under Uncertainty: Stochastic Optimization
Generative Reliability-Based Design Optimization Using In-Context Learning Capabilities of Large Language Models
🤖 AI Summary
Engineering design optimization under reliability constraints faces challenges due to expensive high-fidelity reliability evaluations and difficulty in explicitly encoding complex constraint logic.
Method: This paper proposes an LLM-driven hybrid optimization framework that integrates the in-context learning capability of DeepSeek-V3 into a metaheuristic search process. Through prompt engineering, the LLM generates candidate designs satisfying reliability constraints implicitly; a Kriging surrogate model further reduces computational cost by approximating high-fidelity reliability assessments. The framework enables constraint-aware, intelligent iterative optimization without explicit constraint modeling.
Contribution/Results: This work pioneers the use of LLMs’ generative reasoning in reliability-based design optimization. Evaluated on three canonical engineering cases, the method matches the convergence speed of conventional genetic algorithms while substantially improving feasibility rates and design efficiency—offering a novel paradigm for reliability optimization under data scarcity and costly evaluations.
Technology Category
Application Category
📝 Abstract
Large Language Models (LLMs) have demonstrated remarkable in-context learning capabilities, enabling flexible utilization of limited historical information to play pivotal roles in reasoning, problem-solving, and complex pattern recognition tasks. Inspired by the successful applications of LLMs in multiple domains, this paper proposes a generative design method by leveraging the in-context learning capabilities of LLMs with the iterative search mechanisms of metaheuristic algorithms for solving reliability-based design optimization problems. In detail, reliability analysis is performed by engaging the LLMs and Kriging surrogate modeling to overcome the computational burden. By dynamically providing critical information of design points to the LLMs with prompt engineering, the method enables rapid generation of high-quality design alternatives that satisfy reliability constraints while achieving performance optimization. With the Deepseek-V3 model, three case studies are used to demonstrated the performance of the proposed approach. Experimental results indicate that the proposed LLM-RBDO method successfully identifies feasible solutions that meet reliability constraints while achieving a comparable convergence rate compared to traditional genetic algorithms.
Problem
Research questions and friction points this paper is trying to address.
Leveraging LLMs for reliability-based design optimization
Combining in-context learning with metaheuristic algorithms
Reducing computational burden via LLMs and surrogate modeling
Innovation
Methods, ideas, or system contributions that make the work stand out.
LLMs enable in-context learning for design optimization
Kriging modeling reduces computational burden effectively
Prompt engineering enhances dynamic design generation
🔎 Similar Papers
No similar papers found.
