Integrating a Modified Desirability Function Approach and Ant Colony Strategy to Solve Two-stage Multi-Response Optimization Problem

Evaluation of a product or process involves simultaneous study of several product quality characteristics or responses. The ultimate goal of manufacturing process optimization is to determine the settings of the process variables such that the best combination of the responses is obtained. In case of multiple processing stages, determining optimal process condition is a critical and difficult task. Interdependency between stages, multiple responses in more than one stage and correlation between responses add up to the complexity of the optimization problem. Individual optimal at each stage will not lead to overall best condition for a multistage and multiple response problem. In this paper, an integrated solution approach is recommended and used to solve a two-stage manufacturing process optimization problem. A modified desirability function and ant colony-based search strategy is used to derive the best solution. The case study selected is a two-stage machining of an automobile engine cylinder liner.