Predicting future energy demand will allow for better planning and operation of electricity providers. Suppliers will have an idea of what they need to prepare for, thereby preventing over and under-production. This can save money and make the energy industry more efficient. We applied a multiple regression model and three Convolutional Neural Networks (CNNs) in order to predict Florida’s future electricity use. The multiple regression model was a time series model that included all the variables and employed a regression equation. The univariant CNN only accounts for the energy consumption variable. The multichannel network takes into account all the time series variables. The multihead network created a CNN model for each of the variables and then combined them through concatenation. For all of the models, the dataset was split up into training and testing data so the predictions could be compared to the actual values in order to avoid overfitting and to provide an unbiased estimate of model accuracy. Historical data from January 2010 to December 2017 were used. The results for the multiple regression model concluded that the variables month, Cooling Degree Days, Heating Degree Days and GDP were significant in predicting future electricity demand. Other multiple regression models were formulated that utilized other variables that were correlated to the variables in the best-selected model. These variables included: number of visitors to the state, population, number of consumers and number of households. For the CNNs, the univariant predictions had more diverse and higher Root Mean Squared Error (RMSE) values compared to the multichannel and multihead network. The multichannel network performed the best out of the three CNNs. In summary, the multichannel model was found to be the best at predicting future electricity demand out of all the models considered, including the regression model based on the datasets employed.

An agent-based modeling framework is developed and employed to replicate the interactions among urban farms. The objectives are to efficiently manage an urban farm’s food, energy, and water resources, decrease food waste, and increase the food availability for the local community. A case study of eleven farms was investigated in Vancouver, Canada to study the linkages between the resources in the urban food, energy, and water nexus. Each urban farm in the simulation belonged to a community microgrid generating electricity from solar and wind. The local farms aimed to provide fresh produce for their respective local communities. However, at some points, they lacked supply, and at other points, there was excess supply, leading to food waste. Food waste can be converted into fertilizers or bioenergy. However, an alternative solution must be employed due to the natural resources required for production, efficiently managing resources, and adhering to sustainability guidelines. In this paper, an optimization framework was integrated within the agent-based model to create a micro supply chain. The supply chain directly linked the producers with the consumers by severing the links involved in a traditional food supply. Each urban farm in the study collaborated to reduce food wastage and meet consumer demands, establishing farmer-to-farmer exchange in transitional agriculture. The optimization-based micro supply chain aimed to minimize costs and meet the equilibrium between food supply and demand. Regular communication between the farms reduced food waste by 96.9% over 16 weeks. As a result, the fresh food availability increased for the local community, as exemplified by the consumer purchases over the same period. Moreover, the simulation results indicated that the renewable energy generation at the community microgrids aided in the generation of 22,774 Mwh from solar and 2568 Mwh from wind. This has the potential to significantly reduce CO₂ emissions in areas that heavily rely on non-renewable energy sources.

This paper presents a framework for effective last-mile delivery in underdeveloped megacities by combining social media, machine learning, and reinforcement learning. Leveraging a Graph Convolutional Networks and a Long Short-Term Memory model for traffic prediction, the framework incorporates multimodal data sources, such as social media sentiment analysis, to provide real-time insights into traffic dynamics. By framing the delivery problem as a Markov Decision Process, reinforcement learning dynamically adapts routing decisions to obtain delivery efficiency, reduce delays, and minimize fuel consumption. A case study in Bogotá demonstrates the framework’s effectiveness in mitigating urban traffic challenges. This work highlights the transformative potential of integrating adaptive learning technologies to address urban logistics’ environmental, economic, and operational complexities. Future research explores advanced methodologies, including multi-agent systems and transformer-based architectures, to further enhance scalability and adaptability in dynamic urban environments.

It is well accepted that the pervaporation (PV) permeation was affected by the physicochemical properties of permeant-membrane materials and process conditions. Given many experimental data of PV, a predictive model on PV permeation based on their physicochemical properties and process conditions can be constructed. This study proposes a machine learning approach in terms of artificial neural network (ANN) to predict the permeation flux of PV, as a function of physicochemical properties of permeant-membrane material and process conditions. A large dataset was assembled from the literature and was divided into a training subset and a testing subset. The output variable was the PV flux, while the input variables were the physicochemical properties of permeant-membrane and the process conditions which were considered to affect the PV flux. Two types of inputs were evaluated: regular variables (Type I) and dimensionless groups derived from these regular variables (Type II). Several neural network architectures were evaluated. The best predictive performance for Type I inputs was achieved with a deep neural network consisting of two layers, each with 7 neurons. For Type II inputs, the optimal architecture was a shallow neural network with a single layer containing 6 neurons. The correlation coefficients (R) during model training for Type I and Type II were 0.93674 and 0.88332, respectively, while the root mean square errors (RMSE) were 42.2558 and 38.0766, respectively. The extent of dependency of output on input variables was determined using Garson's equation. It was found that the most affecting physicochemical properties on the PV permeation flux were glass transition temperature (T_gm), solubility difference of two different permeants $\left(\Delta So{l}_{ij}\right)$, and molar volume of permeant (v_i), consecutively. Whereas the operating condition that dominantly affects the PV permeation flux was permeate pressure (P_p).

Artificial intelligence (AI) is continuing to gain traction as technology improves and there is increased data collection. AI can assist society by predicting future phenomena and thus this leads to more efficient processes, increased profits, and lower costs. The evolving world environment is leading to increased pressure for food, energy and water resources and it becomes increasingly important to manage these resources. AI can be utilized in the form of the internet of things (IoT), which are sensors and devices that collect data, to help manage natural resources. In this paper, several machine learning models are employed to predict crop yield. Crop yield is an important factor for many countries and can be an indicator of food security. The ability to predict crop yield is of increased importance as it can aid policymakers in planning future food supply and can aid in the development of a smart irrigation system. A case study of sugarcane in the state of Florida is utilized, along with climate and soil variables. As a follow up to the results of the machine learning models on crop prediction, this paper introduces an optimization model that develops the smart irrigation system. The model chooses the optimal amount of irrigation and fertilizer needed to maximize crop yield while also minimizing costs, water, and energy resources. The machine learning models identify the most important variables needed for crop prediction and can be applied to other crops in other regions of the world.