Quantum-AI Hybrid Algorithms for Solving Large-Scale Engineering Problems Using LLM-Driven Optimizations

Major engineering challenges are hard to solve computationally because they are very complex, have a high number of dimensions and they are non-linear. The paper introduces a new Quantum-AI Hybrid Algorithmic Framework that unites quantum computing and artificial intelligence, mainly Large Language Models (LLMs), to address such issues efficiently. It uses quantum processes to help with probabilistic sampling and reducing the search space and it uses LLMs for logical reasoning, generating code and adjusting parameters to fit the requirements of the given problem. Adding knowledge from a specific engineering area into the update stages using LLM prompts helps the suggested system improve both the speed of convergence and scalability compared to AI-only methods. Testing on structural design, power system development and fluid dynamics problems showed that GhostFlow converges faster than other similar software (up to $40 \%$ faster) and results in improved accuracy (up to $25 \%$ better). It highlights how enhancing quantum-AI with LLM intelligence can lead to important improvements in how computing works in the future.