Combination of SKU to POD Assignment for Robotic Mobile Fulfillment Systems (RMFS)

Pratiwi, Dinda Tria (2024) Combination of SKU to POD Assignment for Robotic Mobile Fulfillment Systems (RMFS). Masters thesis, Institut Teknologi Sepuluh Nopember.

Text 6010211021-Master_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 April 2026. Download (1MB) | Request a copy

Abstract

Robotic Mobile Fulfillment Systems (RMFS) merupakan solusi yang diterapkan secara luas di e-commerce warehouses. Komponen utama sistem ini mencakup pods, storage locations, order picking atau replenishment workstations, dan mobile robots. Dalam konteks sistem gudang, pengambilan keputusan melibatkan aspek strategis, taktis, dan operasional. Salah satu keputusan taktis yang signifikan adalah Product assignment. Product assignment merupakan keputusan taktis yang memengaruhi efisiensi pengambilan pesanan. Product assignment melibatkan alokasi SKU ke pod, pod ke zones, dan penyebaran SKU ke banyak pod. Fokus penelitian ini adalah pada alokasi SKU ke pod, di mana alokasi SKU ke pod merupakan proses awal sebelum simulasi. Ada dua tahap dalam alokasi SKU ke pod, yaitu pengelompokan produk dan kombinasi produk. Pengelompokan produk yang tepat dapat meningkatkan efisiensi pengambilan pesanan. Kombinasi produk dalam alokasi pod mengoptimalkan jumlah unit yang diambil oleh pod, mengurangi penggunaan untuk Automated Mobile Vehicles (AMV). Penelitian ini menguji tiga skenario yaitu: Random-Baseline, Class Combination, dan Cluster Combination. Class Combination menggunakan klasifikasi ABC untuk membagi SKU ke dalam kelas berdasarkan metode Pareto: sejumlah persentase SKU berkontribusi pada persentase frekuensi pesanan, sementara Cluster Combination menggunakan dimensi untuk meletakkan produk di dalam pod. Simulasi dilakukan untuk menentukan pile-on terbaik berdasarkan jumlah unit yang diambil per kunjungan pod ke stasiun pengambilan pesanan. Jumlah unit yang lebih tinggi yang dikumpulkan per kunjungan pod berarti lebih sedikit penggunaan mobile robotnya, menghasilkan pile-on yang lebih efisien. Hasil simulasi menunjukkan bahwa skenario terakhir, yaitu Cluster Combination, menghasilkan pile-on terbaik, Dimana jumlah rata-rata unit yang diambil setiap kunjungan podnya lebih baik dibandingkan dengan dua skenario lainnya. Skenario terakhir menunjukkan peningkatan sebesar 7,39% dan 3,28% dibandingkan dengan skenario pertama dan kedua, secara berturut-turut. Hasil ini divalidasi menggunakan ANOVA, mengkonfirmasi bahwa skenario terakhir mengoptimalkan efisiensi pile-on dalam Robotic Mobile Fulfilment System.
===================================================================================================================================
The Robotic Mobile Fulfilment System (RMFS) is a popular solution in e-commerce warehouses. It consists of pods, storage locations, order picking or replenishment workstations, and mobile robots. Decision problems in warehouse systems are strategic, tactical, and operational. One of the decision problems is product assignment. Product assignment is a tactical decision that affects picking efficiency. Product assignment includes allocation of SKUs to pods, pods to zones, and dispersion of SKUs to many pods. The focus of this research is on SKU to pod allocation, where SKU to pod is the initial process before simulation. There are two stages in SKU to pod allocation, namely product grouping and product combination. Appropriate product grouping can also increase picking efficiency. The product combination in the pod allocation optimizes the number of units being picked by the pod, which is pile-on, reducing the need for Automated Mobile Vehicles (AMVs). Three scenarios were tested: Random-Baseline, Class Combination, and Cluster Combination. Class Combination uses ABC classification to divide SKUs into classes based on Pareto’s method: the percentage number of SKUs contributes to percentages of order frequency, while Cluster Combination uses dimensions when putting the product in the pod. Simulation is the major technique for determining the best pile-on based on the number of units picked per pod visit to the pick station. A higher number of units collected per pod visit means fewer mobile robots in travel, resulting in a more efficient pile-on. Simulation findings show that the final scenario, which is Cluster Combination produced the best pile-on, exceeding the average amount of units selected every pod visit in contrast to the other two situations. The final scenario showed a 7.39% and 3.28% improvement over the first and second situations, respectively. These findings were thoroughly tested using one-way ANOVA, confirming that the final scenario optimizes pile-on efficiency in the Robotic Mobile Fulfilment System.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	RMFS, ABC Classification, Combination of SKU in Pod Assignment, Pile-on, Cluster.
Subjects:	H Social Sciences > HF Commerce > HF5485 Warehouses--Management T Technology > T Technology (General) > T57.62 Simulation
Divisions:	Faculty of Industrial Technology and Systems Engineering (INDSYS) > Industrial Engineering > 26101-(S2) Master Thesis
Depositing User:	Dinda Tria Pratiwi
Date Deposited:	14 Feb 2024 15:47
Last Modified:	14 Feb 2024 15:47
URI:	http://repository.its.ac.id/id/eprint/107262

Actions (login required)

View Item