| Abstract [eng] |
Orders picking is one of the key functions of the warehouse. Analysis shows that the most of the picking is done manually. Therefore, orders picking is very labor-intensive process. The aim of this work is to develop an algorithm, which would improve orders picking process in warehouses. By creating batches, this algorithm should shorten the travel distance needed to pick all the items of the orders. Thesis consist of three parts. In the first part, literature analysis is done. The most often used and best performing algorithms for order batching are identified and compared. Also, two benchmark algorithms are selected. In the second part a new algorithm is developed. The new algorithm is population based Genetic algorithm. The performance of the developed algorithm and benchmark algorithms is than measured and compared in the simulated simplified warehouse environment. Results show that new algorithm generates better order batches than benchmark algorithms. However, the Genetic algorithm needs more time to complete the computations. Based on the insights on the results the Genetic algorithm is updated. The real warehouse simulation environment is than configured in the third part. The orders batching algorithm used in this warehouse is identified and simulated. The performance of the Genetic algorithm developed in this work is compared against the algorithm which is used in this warehouse. Results of the experiment show that Genetic algorithm would provide 17% shorter travel distances on average, when compared to current solution used in the warehouse. |
