Robust optimization of rolling parameters of coarse aggregates based on improved response surface method using satisfaction function method based on entropy and adaptive chaotic gray wolf optimization

Abstract Determining the rolling parameter control standard for coarse aggregates by robust optimization is of great significance in reducing the sensitivity to noise in data on site and improving the measurement accuracy and stability of the compaction quality. The existing rolling parameter control standard is mainly determined by field rolling tests, which results in draining human labor and financial resources and makes it difficult to ensure the global optimum. Existing studies on the multiobjective optimization of rolling parameters usually ignore robustness, which in detail is represented by neglecting the overlap width of the rolling strip in the decision variables and rolling efficiency in optimization objectives. To overcome this limitation, a robust optimization design method for the rolling parameters based on the improved response surface method (RSM) using the satisfaction function method based on entropy (SFME) and adaptive chaotic gray wolf optimization (ACGWO) is proposed. First, the mean and standard deviation (SD) of the rolling quality and efficiency are selected as the response objectives, with the rolling thickness, rolling passes, rolling speed, and overlap width as the influencing factors. Then, RSM is used to perform a robust optimization design on the rolling parameters, in which the SFME is used to determine the weight of each response. Finally, an improved ACGWO algorithm with less calculation time, faster convergence speed, and higher accuracy is proposed to establish a robust optimization model of the rolling parameters. The engineering application results indicate that the most robust rolling parameters obtained by the proposed method are applicable and accurate. The method can improve the efficiency of the compaction while maintaining high compaction quality.