LAUSR

A data‐driven evolutionary genetic algorithm is successfully implemented to optimize two objectives simultaneously from the data generated at an integrated steel plant in view of an efficient continuous casting production line. The objective of optimization is to minimize tundish temperature and maximize the slab exit temperature during continuous casting of steel. EvoNN (Evolutionary Neural Network) coupled with predator prey genetic algorithm is used explicitly to achieve the optimum trade‐off between tundish temperature and slab exit temperature through training of 100 data set considering five process variables: slab width, mold water pressure, average mold water flow, casting speed, and mold water inlet temperature. Optimum trade off curve generated between these two conflicting objectives called as pareto front to yield a set of mutually non‐dominated solutions that helps provide an effective guideline for decision making during the slab production in continuous casting process. It is observed that after a particular critical tundish temperature (i.e., 1555 °C), further enhancement of tundish temperature does not influence the temperature profile of a slab coming out from the caster machine. By optimizing the slab exit temperature (i.e., 815 °C), the proper solidification of slab is achieved with a targeted metallurgical length.