Refine
Has Fulltext
- no (4)
Document Type
Language
- English (4)
Is part of the Bibliography
- yes (4)
Keywords
- Deep learning (1)
- Stacking problem (1)
- Warehouse safety (1)
Institute
Recent risk incidents, such as 911 attacks, financial crisis, catastrophic disasters like earthquakes and floods, etc. have added challenges and susceptibility for business operations. In encountering more frequent and unpredictable risks than before, many companies adopt risk management systems and redesign their supply chain (SC) structures for stronger SC resilience as well as quicker recovery speed. In this paper, we present a case study of a large semiconductor packaging and testing company in Taiwan, AKH, in confronting risk incidents and the plans they implemented to strengthen the SC resilience to assure more reliable supply of their products and services to the customers worldwide. AKH executed several de-risking plans, which can be divided into four Rs - resourcefulness, redundancy, robustness and rapidity. Many of the de-risking plans rely on the E-Hub system the company implemented a few years ago for SC information visibility. After carrying out the de-risking plans, AKH found that the suppliers and customers as well as the company have to furnish additional kinds of information to the E-Hub system in order to improve the SC resilience and assess the impacts of risk incidents. The results show improved SC resilience for the case company in encountering risk incidents and much faster recovery speed once an incident occu
Purpose:
Kiva robots system is considered as the 8th generation of warehousing picking system by Amazon. The system utilizes Automated Guided Vehicles to pick up and carry racks to picking stations to realize goods-to-man picking operations and can save 2/3 of manpower in picking operations. One of the ways for efficient operations of the system is to have right assortment of goods on racks to reduce the number of rack move completing customer orders. This research looks into correlations of customer orders to study the goods allocation plan on racks that could effectively reduce the number of rack moves for shorter order completion time and less manpower used in picking operations.
Design/methodology/approach:
This study aims at finding goods allocation plans
considering correlations of customer orders for efficient picking operations. In calculating the correlation between and among orders, we used a priori association rules to compute the support and confidence of order items. We generate three batches of 10,000 orders with VBA, each referring to an EC company. We later implemented ABC classification and association rules to conduct different levels of order analysis. Considering put-to-light picking systems as in the Kiva systems, the correlations of order items and customer orders were computed with R programming software. The performance measure is the number of times to move racks to complete the customer orders. Different replenishment strategies for goods allocation on racks using product correlation values and order quantity ratios were implemented with simulation to find the optimal replenishment strategy.
Findings:
We analyzed nine replenishment strategies and found that the best strategy among the nine could reduce the number of rack moves by 77.9% when compared to a randomly allocated strategy.
Value/Originality:
Despite the growing application of Kiva systems in warehousing operations, there is only few studies look into operational improvement problems. This paper lies its originality in using association rules for rack replenishment strategy for efficient picking operations. The result show tremendous improvement in performance measures, which demonstrates the value of the paper.
Purpose of this paper:
Image processing technology has gained much improvement in recent years. In smart warehousing, most focus is on warehouse operations to improve efficiency and save human labors. However, in put-away, re-warehousing or picking process, after being handled by forklifts or laborers, the stacking of inventory items may change its shape and balance due to deviation, compression, tilting, etc., which may jeopardize the safety of stored goods and even people working in the warehouse. If there is a system that could monitor the stacking of inventory items all the time and could alarm managers should the stacking tilts or deforms to a certain level, inventory safety could be vastly improved. This study is on smart warehouse safety problems emphasizing at goods stacking to avoid the possibility of falling or collapse through image processing via Mask R-CNN. Our experiment is limited to items that form a rectangular block on a pallet, such as wooden bars, cartons or any box-shaped items.
Design/methodology/approach STS:
Mask R-CNN is a deep learning software and can be used in a wide range of applications. In this study, we used Mask R-CNN to detect the images of stacked items to see if the stacking is stable. We took photos from a warehouse and divided them into three classes – training, validation and production. VGG Image Annotator (VIA) was used for object annotation. To detect the tilt degree of stacked items, we derived formulation from book- stacking problems and retrieved dot coordinates of square frames encompassing objects in an image from the result of Mask R-CNN for calculation.
Findings:
Because there is a limited number of photos for training, validation and production, we applied image enhancement techniques such as flipping, affinity, multiplication, etc. to increase the total number of images to avoid over-fitting during the training process. The trained Mask R-CNN network demonstrated its capability in framing objects and after retrieving the dot coordinates of the frames, it could compute the degree of tilt as a reference for block stacking alarm.
Value:
The paper presents a relatively rare study of applying deep learning techniques on images as applied to material stacking safety problems. It contributes to a special smart warehousing domain. The same technique can be applied to inventory counting and help improve labor-saving in counting and in locating items in a warehouse.
Research limitations/implications:
Due to the limited number of images and restricted angles of taking the images, Mask R- CNN may not be so capable in framing objects of special angles of images. More images with diverse angles of them should be obtained for training to improve the capability of the trained network.
Practical implications:
Successful implementation of the technology developed in this study in a warehouse monitoring system can help industries in smart detection of improper stacking to assist human in improving the accuracy of safety judgement and can also save labors in reviewing images or records of a warehouse.
Most research in the past mainly considers AS/RS in distribution environment. In this study, we consider AS/RS in the support of an assembly line having multiple cells. For the coordination between the assembly line and the AS/RS, we implemented metaheuristics with Tabu algorithm to solve the sequencing of tasks having picking and put-away operations in support of a Toyota manufacturing system where production lines are composed of cells to assemble simultaneously a variety of products such as auto parts, electronic devices, etc. In this paper, three Tabu heuristic algorithms were devised with variations in search and exchange. The first algorithm (TI) is a regular Tabu search with special initial solution strategy; the second algorithm (TID) combines TI with divide and conquer search strategy; the third algorithm (TIE) is TI with special exchange strategy. We experimented on an AS/RS – production system having 40 assembly cells and an AS/RS with capacity of 1600 baskets for holding materials for assembly. By making 30 runs for each of the above three algorithms, we found that that TIE performs the best, followed by TID and then TI. It also demonstrates that Tabu search can help the operation of AS/RS by reducing 5-30% of shop floor assembly time while TIE can obtain a sequence solution in the best possible time by saving 54.9% of the computation time over TI.