INTERNATIONAL SCIENTIFIC JOURNAL OF ENGINEERING AND TECHNOLOGY (ISJET)

A Comparison between Profit and Economic Value Added Optimization to Design a Supply Chain Network: A Case Study of Food Supply Chain in Vietnam

2024-06-05T14:31:51+07:00

The research field of food supply chain network design and optimization has expanded significantly. However, most recent studies have focused only on minimizing costs or maximizing profits, neglecting other important financial factors that aﬀect the overall prosperity of the chain. This study generates two scenarios in the design of the supply chain network, comparing the maximization
of profit and Economic Value Added (EVA) to assess their eﬀectiveness in real-world situations. The comparison is based on supplier and potential distribution center selection, along with considerations of production level, production capacity, and the sizes of the plant, distribution centers, and retailers. The methodology considered in this research is based on Mixed-Integer Linear Programming (MILP) under deterministic parameters. The study provides computational results and managerial insights based on a case study of the food supply chain in Southern Vietnam. The findings indicate that the EVA maximization model oﬀers a more precise evaluation of company wealth as compared to the profit maximization model as it can determine more suitable operating supply chain’s decision variables leading to a significant decrease of 11.1% in the invested capital.

A Literature Review of Steering Angle Prediction Algorithms for Autonomous Cars

2024-02-22T15:35:41+07:00

Road tracking is a critical requirement for the development of autonomous cars. It requires the car to continuously navigate within the designated driving area to avoid any deviation. The computation of steering angles is an essential aspect of achieving autonomous driving. Autonomous steering angle techniques, which are essential for enabling road tracking in autonomous cars, are comprehensively reviewed in this paper. Autonomous steering techniques, mainly involving computer vision methods and end-to-end deep learning approaches, are currently receiving considerable attention. The primary objective of this paper is to identify and reimplement state-of-the-art models in end-to-end deep learning approaches within practical scenarios. We carry out a performance evaluation of each model utilizing real-world tests using scale model cars. Furthermore, we offer perspectives on potential avenues for future research and applications. These may include adaptive modifications to dynamic road conditions, the creation of more effective real-time decision-making algorithms, or the investigation of applications in intricate traffic situations.

A Transfer Learning-based Deep Convolutional Neural Network Approach for White Shrimp Abnormality Classifcation

2024-09-27T09:16:33+07:00

Shrimp transportation frequently leads to product damage, necessitating a sorting system to identify and remove compromised shrimp prior to processing. This research aims to develop a transfer learning-based deep convolutional neural network system capable of accurately categorizing shrimp into seven classes: Complete body, crunched head, head loss, head loss with remaining chin, cut tail, torn in half, and total crunched. A dataset comprising 405 color shrimp images, each with dimensions of 1,920 x 1,080 pixels, was augmented using geometric transformations to expand the dataset to 6,480 images. These augmented images were then employed to train four state-of-the-art transfer learning-based models (NasNetLarge, InceptionResNetV2, EfcientNetV2L, ConvNeXtXLarge) from Keras Applications. These models also were subsequently compared to a baseline CNN. Results demonstrate that the ConvNeXtXLarge model outperformed the others, achieving the highest accuracy (95%), precision (0.96%), recall (0.95%), and F1-score (0.95%), underscoring its superiority in shrimp damage classification. An analysis of misclassifications revealed potential confusion between certain damage classes, suggesting areas for future refinement to enhance the model’s ability to diﬀerentiate between similar types of damage.

Enhancing Warehouse Management with AI and Computer Vision: A Case Study in a Logistics Service Company

2024-06-21T16:14:29+07:00

In the evolving landscape of warehouse management in Industry 4.0, this paper explores the convergence of Artificial Intelligence (AI) and Computer Vision (CV) for inventory tracking and stock registration. Conducted in collaboration between SIIT and KNS, a logistics
service company specializing in warehousing, the study introduces a framework that optimizes image capture conditions through real-time analysis of gyroscope values, distinguishing mobile phone movement from stationary states. Additionally, an object detection model using the YOLOv8 algorithm achieves 83% accuracy in label detection and 75% in box detection within a curated dataset. The research highlights the successful development of the phone motion detection model and Optical Character Recognition (OCR) integration. This framework promises to advance warehouse management systems by addressing current limitations with a comprehensive, efficient, and user-friendly solution.

The Development of Smart Farming System for Sea Lettuce Cultured Process

2024-07-30T13:04:02+07:00

Smart farming is a progressive development that leverages information and communication technology in machinery, equipment, and sensors within network-based hi-tech farm supervision cycles. At the forefront of this is the Internet of Things (IoT), a new technology that enables remote device connectivity for smart farming. This paper sets out to transform the seaweed farming process into an intelligent farm system with water oxygen control and environmental monitoring, all powered by IoT technology. The IoT gateway, built with Node-RED, facilitates device connectivity through an Application Programming Interface, with a Node-RED Dashboard providing real-time data in a graphical format. Experiments have shown that the oxygen control system can be operated using two valves, and a valve system can be controlled through a web application, Node-RED, using IoT technology and Internet network connectivity. The embedded system's percentage error is configured as not exceeding 5 percent for the water temperature, the air temperature, and the humidity measurement. The Node-RED Dashboard displays real-time data, including node status of valve control oxygen, water temperature, air temperature, humidity, and the switch of the valves control oxygen. The data can be transmitted via a wireless system of devices based on the IoT concept.

The Morphology Associated with Harvesting Stages of Siam Red Ruby Pumelo (Citrus grandis)

2024-11-21T10:28:55+07:00

This study aims to evaluate fruit morphological characteristics and potential production of ‘Siam Red Ruby’ pumelo from Pak
Panang district, Nakhon Si Thammarat Province compared between normal and senescence fruit peel. A three-year-old tree was selected to evaluate fruit morphological characteristics and a harvesting date or time at 90, 120, 150, 180, and 210 days after fruit Set (dAfS). The result showed anatomical association to the senescence of ‘Siam Red Ruby’ pummelo using the peel of normal and senescence zones using compound microscopy. The ‘Siam Red Ruby’ pummelo was shown a green color inside a peel 90 days after fruit setting and in a peel showed a trichome or hair covering all the fruit (100%) then decreased during fruit development until 210
days after fruit setting showed the hair cover 20%in all fruits compared with others stage. during the maturity stage, the pulp was yellow-colored from 90 to 120 days after the fruit set. from 150 days after the fruit set, the pulp turned red gradually. In addition, the structure of the peel was changed during senescence. The first day of the peel or normal zone showed high firmness and peel green color. In the senescence zone, the peel changed to a yellow color. In addition, Titratable Acidity (TA) decreased from 90 to 210 days after fruit setting and related with Total Soluble Solid (TSS) increased from mature fruit (90 days after fruit setting) to ripening fruit (210 days after fruit setting). In reﬂective sheet treatment, clouds induce the red color in the pulp of ‘Siam Red Ruby’ pumelo.