Stack Shuffling Optimization of Steel Bars by Using Genetic Algorithms

A steel plant company producing steel bars has large assortment of the end products, with similar appearance and attributes. The steel bars are stored on the floor in a stacking frame. For the order picking of steel bars, an overhead crane is used for reshuffling all the necessary steel bars to get access to the required product. While the production schedule allows for anticipating the storage occupancy, a stochastic transport arrival prevents optimal product stacking for efficient order-picking operation. Due to this, any order-picking sequence may require reshuffling of the stacked material, which increases working cost, order-picking times, and complicates material tracking. This paper presents a method for minimizing the order-picking times by overhead crane movements through proper reshuffling of the steel bars. Similar research was done on container yard pre-marshalling and reshuffling problem, while the presented approach handles with the special situation in the steel plant. Various optimization approaches including linear programming, simulated annealing, taboo search, branch and bound and genetic algorithms were used by researchers to solve comparable problems. The proposed method for solving the specific problem of reshuffling steel bars uses genetic algorithms to find a feasible solution in real-time. The proposed solution reduces intralogistics cost and increases order-picking efficiency.