Automated generation of executable test suites from natural-language requirements remains challenging due to linguistic ambiguity and sensitivity of generative models to decoding and training hyperparameters. This paper introduces a hierarchical, multi-level evolutionary framework that treats model hyperparameters and decod- ing strategies as upper-level decision variables and em- ploys lower-level fitnesses that directly measure test-quality objectives (structural coverage, semantic diversity, redun- dancy, and runtime efficiency). The approach integrates retrieval-augmented grounding, surrogate-assisted preselec- tion, lightweight LoRA adaptation and optional HIL evalu- ation. Empirical evaluation on PURE, PROMISE exp and FR NFR benchmarks (repeated runs, n = 10; paired two- sided t-tests, α = 0.05) shows consistent gains: on PURE mean code coverage reaches 82.4% (vs. 75.1% for Bayesian optimisation and 68.9% for random search) with 145 unique scenarios and modest runtime overhead (≈58.3 s, ≈6% above Bayesian). Ablations confirm component effects (e.g., removing diversity reduces unique scenarios ≈18%; disabling the surrogate increases wall-clock ≈42%; dis- abling RAG drops grounded consistency ≈12%). Results indicate that co-optimising hyperparameters for explicit test-quality metrics, together with grounding and realistic execution, yields more useful, executable test suites. Future work will explore adaptive objective weighting, transfer warm-starts and probabilistic surrogates.