Muhammad Asrol

The cold chain industry in Indonesia is experiencing rapid growth, especially for perishable products such as beef. Ensuring product quality during distribution requires accurate monitoring of storage environmental factors, including temperature, humidity, and gas exposure. This study aims to develop an IoT-based quality monitoring system for perishable beef products and implement a machine learning approach for quality prediction. An IoT-enabled e-Sense device was developed to collect real-time environmental parameters and RGB colour information as quality parameter from tenderloin beef samples. The collected data were analysed using three regression-based machine learning algorithms: Random Forest (RF), Decision Tree (DT), and Support Vector Regression (SVR). Data preprocessing and hyperparameter tuning were applied to improve model performance. The results show that SVR consistently outperformed RF and DT in predicting RGB colour in presenting beef quality based on prescribed parameters. SVR achieved an R² of 0.973 for RED and 0.992 for both GREEN and BLUE channels. These findings confirm the effectiveness of integrating IoT technology with machine learning for real-time perishable products quality prediction. This research contributes to combining real-time multi-sensor IoT data with regression-based models to provide improved continuous quality monitoring compared to previous single-parameter or offline approaches.

Meat quality prediction is an important component of data-driven supply chain management. However, the uncertain and nonlinear nature of the data often reduces the accuracy of prediction models. This study aims to identify the factors influencing meat quality prediction accuracy and to develop an intelligent inference system based on the Interval Type-2 Takagi-Sugeno-Kang (IT2TSK) Fuzzy Inference System (FIS) to handle uncertainty and improve prediction performance. The IT2TSK model was built by automatically generating a rule base and membership functions (MFs) using the mini-batch K-means (MBK) algorithm and Gaussian MFs, which were then optimized in two stages: a two-level Grid Search (GS) for parameter exploration and a Genetic Algorithm (GA) for adaptive tuning. The evaluation was conducted on 12 meat-cut datasets based on electronic nose (e-nose) sensor signals. The results show that the optimized IT2TSK model achieved an average coefficient of determination (R²) of 0.9844, mean squared error (MSE) of 0.0014, and mean absolute error (MAE) of 0.0236, indicating high prediction accuracy and stability compared to conventional fuzzy systems. This model can balance interpretability, uncertainty handling, and prediction accuracy, and thus, has the potential to be applied to a real-time quality monitoring system for perishable products.

Fat-filled powder has the potential to substitute milk in meeting the nutritional needs of the community, but its product quality remains unstable during continuous production processes. A key challenge in fat-filled powder (FFP) production is the difficulty in quality monitoring, which is influenced by various uncertainty factors that affect product quality. Machine learning can be implemented for quality monitoring system, but the imbalanced data conditions require the development of algorithms with optimal performance. This study aims to design a quality monitoring system for FFP using a machine learning model under imbalanced dataset conditions and the influence of other uncertainty factors. A Random Forest (RF) machine learning model was developed for monitoring FFP quality. In the context of imbalanced datasets, the model was optimized through various scenarios, including data splitting for training and testing, as well as the Synthetic Minority Oversampling Technique (SMOTE) and Distribution Optimally Balanced – Stratified Cross Validation (DOB-SCV) schemes. The results showed that the SMOTE model achieved the best performance in terms of accuracy, precision, and recall with scores of 99.67%, 99.79%, and 99.24%, respectively, on the testing data. Statistically, the RF model with the SMOTE data manipulation scenario also showed significant differences compared to the DOB-SCV model and the traditional data splitting approach. The quality monitoring model for FFP developed in this study can be implemented in the dairy industry, offering more stable, accurate quality monitoring predictions that align with real conditions, helping to avoid quality uncertainties during the production process. The implementation of this model in the industry has the potential to facilitate a broader, more transparent, and optimized product quality evaluation process, which can also be conducted in real time under continuous production conditions.

Indonesia’s sugar industry has yet to become self-sufficient in sugar production. This is due to the unpredictable and fluctuating relationship between yield (sugar content (%)) and sugarcane productivity (Ton/Ha) in all state-owned and private Indonesian sugar mills. As a result, Indonesia’s domestic sugar consumption is still balanced by sugar imports. This study aimed to identify the main criteria and predict sugarcane yield and productivity using vegetative growth indicators in sugarcane cultivation using data science and a machine learning technique based on SVR and RF. The study found that the essential features for predicting sugarcane yield are clear juice, Pol, purity, Brix, and maturity factor, whereas the number of stems, stem height, stem weight, rainfall, and juring factor are important for predicting sugarcane productivity. The best model to predict sugarcane yield (%) was generated using RF with an average absolute error rate of 0.074% and accuracy in predicting yield with an average absolute percentage error of 0.010% and a sugarcane yield prediction error rate of 0.129%. The best sugarcane productivity prediction model was generated using SVR with an average absolute error rate of 0.051 tons/ha and accuracy in forecasting productivity with an average absolute percentage error of 0.001% and a sugarcane productivity prediction error rate of 0.058 tons/ha. This model may be used to optimize sugar cane cultivation and harvesting times, resulting in increased productivity and yields, which benefits corporate performance and increases national sugar output.

Energy storage systems are becoming increasingly important as more renewable energy systems are integrated into the electrical (or power utility) grid. Low-cost and reliable energy storage is paramount if renewable energy systems are to be increasingly integrated into the power grid. Lead-acid batteries are widely used as energy storage for stationary renewable energy systems and agriculture due to their low cost, especially compared to lithium-ion batteries (LIB). However, lead-acid battery technology suffers from system degradation and a relatively short lifetime, largely due to its charging/discharging cycles. In the present study, we use Machine Learning methodology to estimate the battery degradation in an energy storage system. It uses two types of datasets: discharge condition and lead acid battery data. In the initial analysis, the Support Vector Regression (SVR) method with the RBF kernel showed poor results, with a low accuracy value of 0.0127 and RMSE 5377. On the other hand, the Long Short-Term Memory (LSTM) method demonstrated better estimation results with an RMSE value of 0.0688, which is relatively close to 0.

Energy storage systems are becoming increasingly important as more renewable energy systems are integrated into the electrical (or power utility) grid. Low-cost and reliable energy storage is paramount if renewable energy systems are to be increasingly integrated into the power grid. Lead-acid batteries are widely used as energy storage for stationary renewable energy systems and agriculture due to their low cost, especially compared to lithium-ion batteries (LIB). However, lead-acid battery technology suffers from system degradation and relatively short lifetime, largely due to its charging/discharging cycles. In the present study, we use Machine Learning methodology to estimate the battery degradation in an energy storage system. It uses two types of datasets: discharge condition and lead acid battery data. The use of the Long Short-Term Memory (LSTM) model algorithm for estimating battery Remaining Useful Lifetime (RUL) is a promising development that has the potential to significantly improve battery performance/durability and reduce maintenance costs. The initial State of Health (SOH) and predicted SOH values provide insights into battery aging and RUL estimation for lead-acid batteries in the renewable energy systems for agricultural purposes.

The supply chain for perishable products faces significant challenges in monitoring and maintaining product quality. These products are particularly vulnerable to environmental dynamic conditions and variations in distribution and transportation. To address these challenges, leveraging the Internet of Things (IoT) and quality inference techniques during transportation can provide valuable insights for both consumers and producers. The objective of the research is to develop a model for inferring the quality of perishable products using an IoT sensor dataset to monitor perishable product quality continuously. This research applied a hybrid approach combining a Fuzzy Inference System (FIS), clustering models, and genetic algorithms to infer the product quality during supply chain distribution with IoT sensors. The result shows that the hybrid FIS model, which employs Gaussian membership functions and fuzzy c-means clustering for rule generation, achieves a high accuracy with an R²: 0.873. This research contributes to improving the model by employing genetic algorithms in optimizing the inference model by activating only five out of seven rules. The model optimization achieves optimal computation time while aiming to preserve model accuracy. However, test results indicate that the combination of rules has not yet significantly enhanced the model's accuracy, though it holds potential for future development. Doi: 10.28991/ESJ-2025-09-01-027 Full Text: PDF

In the fiercely competitive telecommunications industry of Indonesia, the significance of innovation and perpetual improvement cannot be underestimated. It is imperative for businesses to explore the potential of sustainable supply chain management, as it allows for the integration of environmental, economic, and social dimensions. By implementing a comprehensive and thorough approach, this study offers a theoretical framework for measuring sustainable supply chain performance. Through qualitative and quantitative methods, you'll be able to identify the attributes that greatly impact the success of your sustainable supply chain management. On this research Analytic Hierarchy Process (AHP) utilize to define Consistency Ratio (CR) and weight between dimension using SuperDecisions software. Then, Multidimensional Scaling (MDS) is utilized to measure performance of supply chain sustainability. At the same time, a stress value also measured to indicate the sustainability measurement of one specific dimension is sufficient in accordance with existing condition or not. Next, to better indicate which attributes are most influential on improving dimensional sustainability performance, it is essential to conduct a sensitivity analysis and Montecarlo analysis for each attribute. In the end, this research successfully determined which attributes that influence sustainable supply chain management and measure the value of supply chain sustainability in the telecommunications industry in Indonesia.

Industry 4.0 adoption in industry and business has grown extensively. Supply chain had an opportunity to adopt Industry 4.0 in improving flexibility, efficiency, and response to consumer demands. Considering Industry 4.0, adoption in supply chain operations must be realized as an opportunity and challenge for practitioners. Hence, this study conducted a systematic literature review with a qualitative and quantitative approach to formulate supply chain 4.0 dimensions and operations in assessing readiness for adopting industry 4.0. A Preferred Reporting Item for Systematic Review and Meta-analyses (PRISMA) and bibliometric analysis are performed to analyze the literature related to industry 4.0 adoption in supply chain operations. This research combined a qualitative and quantitative analysis in reviewing the literature to deliver a broader benefit of Industry 4.0 in supply chain operations. This study found 525 papers related to Industry 4.0. Moreover, only twenty papers discussed supply chain specifically. Literature analysis using quantitative and qualitative approach showed that industry 4.0 adoption in the supply chain is limited to the maturity model with digital transformation. This study proposed five comprehensive dimensions to capturing supply chain operations: technology and information technology (IT) infrastructure, supply chain integration and coordination, manufacturing operations and inventory, leadership and human resources, and sustainability. The proposed dimensions for supply chain operations are based on a conceptual analysis of literature. Further verification and empirical analysis is needed to gain deeper insights. While digital technology and transformation are the main focus of Industry 4.0 adoption in many sectors, supply chain management requires broader applications as it involves various sectors and stakeholders. To further analyze this, a readiness and maturity model needs to be developed to assess the adoption of Industry 4.0 in the supply chain, taking into consideration the proposed dimensions.