Analysis of the Challenges of Artificial Intelligence of Things (AIoT) for the Smart Supply Chain (Case Study: FMCG Industries).

In today's competitive world, supply chain management is one of the fundamental issues facing businesses that affects all an organization's activities to produce products and provide services needed by customers. The technological revolution in supply chain logistics is experiencing a significant wave of new innovations and challenges. Despite the current fast digital technologies, customers expect the ordering and delivery process to be faster, and as a result, this has made it easier and more efficient for organizations looking to implement new technologies. "Artificial Intelligence of Things (AIoT)", which means using the Internet of Things to perform intelligent tasks with the help of artificial intelligence integration, is one of these expected innovations that can turn a complex supply chain into an integrated process. AIoT innovations such as data sensors and RFID (radio detection technology), with the power of artificial intelligence analysis, provide information to implement features such as tracking and instant alerts to improve decision making. Such data can become vital information to help improve operations and tasks. However, the same evolving technology with the presence of the Internet and the huge amount of data can pose many challenges for the supply chain and the factors involved. In this study, by conducting a literature review and interviewing experts active in FMCG industries as an available case study, the most important challenges facing the AIoT-powered supply chain were extracted. By examining these challenges using nonlinear quantitative analysis, the importance of these challenges was examined and their causal relationships were identified. The results showed that cybersecurity and a lack of proper infrastructure are the most important challenges facing the AIoT-based supply chain.

Optimizing cropping pattern to improve the performance of irrigation network using system dynamics-Powell algorithm.

One of the strategies for agricultural development is the optimal use of irrigation and drainage networks, which leads to higher productivity and economic benefits. In this regard, quantitative and qualitative studies of drainage water from networks are essential for efficient water management. In the present study, we develop a model using a system dynamics approach to simulate the cropping pattern of an irrigation and drainage network as well as the discharge and salinity of drainage water from network farms. We apply the Powell algorithm to optimize the economic profitability of cultivated crops by considering the salinity and discharge of drainage water from the fields. With three aims, i.e., (1) maximizing benefit-cost ratio, (2) minimizing drainage water salinity and discharge of network, and (3) economic and environmental considerations simultaneously, the optimization of cropping pattern within the Kosar irrigation and drainage network is performed. Results based on five consecutive years under different scenarios showed that some crops, such as watermelon, are not economically recommened for production due to high costs, water consumption, and low selling price causes environmental pollution. On the other hand, wheat, grain maize, silage maize, sorghum, and alfalfa have different conditions, and their production is suitable by considering all scenarios. By comparing with experimental data, we find that the proposed model is accurate to simulate and optimize the irrigation network and to detect its cropping pattern.

Evaluation of geostatistical techniques and their hybrid in modelling of groundwater quality index in the Marand Plain in Iran.

In many parts of the world, groundwater is considered as one of the main sources of urban and rural drinking water. Over the past three decades, the qualitative and quantitative characteristics of aquifers have been negatively affected by different factors such as excessive use of chemical fertilizers in agriculture, indiscreet, and over-exploitation use of groundwater. Therefore, finding the effective method for mapping the water quality index (WQI) is important for locating suitable and non-suitable areas for urban and rural drinking waters. In the present paper, the best method to estimate the spatial distribution of WQI was assessed using the inverse distance weighted, kriging, cokriging, geographically weighted regression (GWR), and hybrid models. Creating hybrid models can increase modeling capabilities. Hybrid methods make use of a combination of estimated model capabilities. In addition, to improve the results of cokriging, GWR, and hybrid methods, the auxiliary parameters of land slope, groundwater table, and groundwater transmissibility were used. In order to assess the proposed methodology, 11 qualitative parameters obtained from 63 observation wells in Marand Plain (Iran) were utilized. Four statistical measures, namely the root mean square error (RMSE), the mean absolute error (MAE), the Akaike coefficient (AIC), and the correlation coefficient (R2) along with the Taylor diagram, have been done. Classification of the WQI index showed that the quality of a number of 1, 27, 18, and 17 wells was, respectively, in excellent, good, moderate, and poor grades. The results of modeling the WQI index based on IDW, kriging, cokriging, GWR, and hybrid methods showed that the best estimate of WQI was obtained by using hybrid GWR-kriging method with three input parameters of land slope, groundwater table, and groundwater transmissibility. Therefore, hybrid kriging and GWR methods have been fairly well able to simulate the WQI index.