CPROO_Data

The draw ratios allocation in drawing process is critical to the final performance of the carbon fiber product. Due to the stochastic nature of component fluxes and concentrations variations, draw ratios exhibit uncertainty, leading to frequent parameter fluctuations. Therefore, this paper investigates the robust operation optimization problem of draw ratios allocation, formulating a model to minimize linear density and maximize strength. To solve this model, a causal forest-guided partitioning for robust operation optimization (CPROO) algorithm is proposed. The CPROO uses perturbations on individual decision variables to form treatment groups from an initial control population, enabling causal forests to infer the overall causal effects of each decision variable on the objectives. Subsequently, the computed mean conditional average treatment effects (CATEs) are clustered via K-Means to partition the decision variables into weakly robustness-related variables and highly robustness-related variables. Then, a differentiated optimization strategy is performed on these two variable groups, using an external archiving mechanism to efficiently search for the robust Pareto optimal set. Finally, adaptive utopian point-based decision making is used to determine the optimal setpoint for draw ratios. Experimental results validate the effectiveness of CPROO in benchmark problems and its practical applicability to the ROOP of draw ratios allocation.