Robust Optimal Desirability Approach for Multiple Responses Optimization with Multiple Productions Scenarios

An optimal desirability function method is proposed to optimize multiple responses in multiple production scenarios, simultaneously. In dynamic environments, changes in production requirements in each condition create different production scenarios. Therefore, in multiple production scenarios like producing in several production lines with different technologies in a factory, various fitted response models are obtained for each response according to their related conditions. In order to consider uncertainty in these models, confidence interval of fitted responses has been defined in the proposed method. This method uses all values in the confidence region of model outputs to define the robustness measure. This method has been applied on the traditional desirability function of each scenario in order to get the best setting of controllable variables for all scenarios simultaneously. To achieve this, the Imperialist Competitive Algorithm has been used to find the robust optimal controllable factors setting. The reported results and analysis of the proposed method confirm efficiency of the proposed approach in a dynamic environment.