<p>Engineering optimization tasks often require tens of thousands of real function evaluations to obtain the optimal solution, which incurs significant computational costs and poses challenges in meeting the requirements of practical engineering optimization problems. To maintain a high level of optimization performance under limited computational budgets, this study proposes a Surrogate-Assisted Grey Wolf Optimization framework based on Query by Committee (SAGWO-QBC). The multi-strategy optimization algorithm framework includes a search strategy based on the maximum uncertainty criterion, a surround strategy utilizing clustering algorithms, and an attack strategy employing cross-validation to construct the mechanism for adding points. The method named Query by Committee is employed to enable adaptive switching between these strategies, effectively filling the solution space and improving algorithm efficiency. To evaluate the performance of the proposed algorithm framework, extensive tests were conducted across various optimization scenarios, leading to the following conclusions: (1) Performance tests on 15 benchmark functions from CEC2005 and CEC2017 (with dimensions ranging from 10 to 30) demonstrate that under limited computational budgets, the proposed algorithm outperforms six state-of-the-art optimization algorithms in over 80% of test cases. (2) Application of the algorithm to three classic constrained engineering optimization problems from the CEC2020 test set reveals that SAGWO-QBC exhibits strong competitive potential compared to the top performing algorithms in the test set. In addition, the internal parameters and optimization process of the algorithm are analyzed and discussed in detail to further understand its operation mechanism. In summary, the SAGWO-QBC algorithm framework demonstrates outstanding performance, delivering high-quality solutions with low computational cost, making it a promising approach for addressing complex engineering optimization problems.</p>