This paper proposes an optimal route and charging station selection (RCS) algorithm based on model-free deep reinforcement learning (DRL) to overcome the uncertainty issues of the traffic conditions and dynamic arrival charging requests. The proposed DRL based RCS algorithm aims to minimize the total travel time of electric vehicles (EV) charging requests from origin to destination using the selection of the optimal route and charging station considering dynamically changing traffic conditions and unknown future requests. In this paper, we formulate this RCS problem as a Markov decision process model with unknown transition probability. A Deep Q network has been adopted with function approximation to find the optimal electric vehicle charging station (EVCS) selection policy. To obtain the feature states for each EVCS, we define the traffic preprocess module, charging preprocess module and feature extract module. The proposed DRL based RCS algorithm is compared with conventional strategies such as minimum distance, minimum travel time, and minimum waiting time. The performance is evaluated in terms of travel time, waiting time, charging time, driving time, and distance under the various distributions and number of EV charging requests.

Electric vehicles (EVs) parking lots are representing significant charging loads for relatively a long period of time. Therefore, the aggregated charging load of EVs may coincide with the peak demand of the distribution power system and can greatly stress the power grid. The stress on the power grid can be characterized by the additional electricity demand and the introduction of a new peak load that may overwhelm both the substations and transmission systems. In order to avoid the stress on the power grid, the parking lot operators are required to limit the penetration level of EVs and optimally distribute the available power among them. This affects the EV owner’s quality of experience (QoE) and thereby reducing the quality of performance (QoP) for the parking lot operators. The QoE is represents the satisfaction level of EV owners; whereas, the QoP is a measurement representing the ratio of EVs with QoE to the total number of EVs. This study proposes a fuzzy logic weight-based charging scheme (FLWCS) to optimally distribute the charging power among the most appropriate EVs in such a way that maximizes the QoP for the parking lot operators under the operational constraints of the power grid. The developed fuzzy inference mechanism resolves the uncertainties and correlates the independent inputs such as state-of-charge, the remaining parking duration and the available power into weighted values for the EVs in each time slot. Once the weight values for all EVs are known, their charging operations are controlled such that the operational constraints of the power grid are respected in each time slot. The proposed FLWCS is applied to a parking lot with different capacities. The simulation results reveal an improved QoP comparing to the conventional first-come-first-served (FCFS) based scheme.

The widespread adoption of electric vehicles (EVs) has entailed the need for the parking lot operators to satisfy the charging and discharging requirements of all the EV owners during their parking duration. Meanwhile, the operational constraints of the power grids limit the amount of simultaneous charging and discharging of all EVs. This affects the EV owner’s quality of experience (QoE) and thereby reducing the quality of performance (QoP) for the parking lot operators. The QoE represents a certain percentage of the EV battery required for its next trip distance; whereas, the QoP refers to the ratio of EVs with satisfied QoE to the total number of EVs during the operational hours of the parking lot. This paper proposes a two-stage fuzzy logic inference based algorithm (TSFLIA) to schedule the charging and discharging operations of EVs in such a way that maximizes the QoP for the parking lot operators under the operational constraints of the power grid. The first stage fuzzy inference system (FIS) of TSFLIA is modeled based on the real-time arrival and departure probability density functions in order to calculate the aggregated charging and discharging energies of EVs according to their next trip distances. The second stage FIS evaluates several dynamic and uncertain input parameters from the electric grid and from EVs to distribute the aggregated energy among the EVs by controlling their charging and discharging operations through preference variables. The feasibility and effectiveness of the proposed algorithm are demonstrated through the IEEE 34-node distribution system.

In this study, we aimed to develop a prediction model of the solid solutions concentration (SSC) and moisture content (MC) in oriental melon with snapshot-type hyperspectral imagery (Visible (VIS): 460–600 nm, 16 bands; Red-Near infrared (Red-NIR): 600–860 nm, 15 bands) using a machine learning model. The oriental melons were cultivated in a hydroponic greenhouse, Republic of Korea, and a total of 91 oriental melons that were harvested from March to April of 2022 were used as samples. The SSC and MC of the oriental melons were measured using destructive methods after taking hyperspectral imagery of the oriental melons. The reflectance spectrum obtained from the hyperspectral imagery was processed by the standard normal variate (SNV) method. Variable importance in projection (VIP) scores were used to select the bands related to SSC and MC. As a result, ten (609, 736, 561, 849, 818, 489, 754, 526, 683, and 597 nm) and six (609, 736, 561, 818, 849, and 489 nm) bands were selected for the SSC and MC, respectively. Four machine learning models, support vector regression (SVR), ridge regression (RR), K-nearest neighbors regression (K-NNR), and random forest regression (RFR), were used to develop models to predict SSC and MC, and their performances were compared. The SVR showed the best performance for predicting both the SSC and MC of the oriental melons. The SVR model achieved a relatively high accuracy with R² values of 0.86 and 0.74 and RMSE values of 1.06 and 1.05 for SSC and MC, respectively. However, it will be necessary to carry out more experiments under various conditions, such as differing maturities of fruits and varying light sources and environments, to achieve more comprehensive predictions and apply them to monitoring robots in the future. Nevertheless, it is considered that the snapshot-type hyperspectral imagery aided by SVR would be a useful tool to predict the SSC and MC of oriental melon. In addition, if the maturity classification model for the oriental melon can be applied to fields, it could lead to less labor and result in high-quality oriental melon production.

In order to decrease fuel consumption and greenhouse gas emissions, electric vehicles (EVs) are being widely adopted as a future transportation system. Accordingly, increasing the number of EVs will mean battery charging will have a significant impact on the power grid. In order to manage EV charging, an intelligent charging strategy is required to prevent the power grid from overloading. Therefore, we propose an optimal energy management algorithm (OEMA) to minimize peak load on a university campus consisting of an educational building with laboratories, a smart parking lot, EVs, photovoltaic (PV) panels and an energy storage system (ESS). Communication networks are used to connect all the system components to a university energy management system (UEMS). The proposed OEMA algorithm coordinates EV charging/discharging so as to reduce the peak load of the building’s power consumption by considering the real-time price (RTP). We also develop a priority determination method for the time allocation of the optimal charging algorithm. Priority is determined by arrival time, departure time, state-of-charge (SOC), battery capacity and trip distance. The performance of the proposed algorithm is evaluated in terms of charging cost and peak load under the real environment of the university engineering building.

Expanding the scale of GPU-based deep learning (DL) clusters would bring not only accelerated AI services but also significant energy consumption costs. In this paper, we propose a cost efficient deep learning job allocation (CE-DLA) approach minimizing the energy consumption cost for the DL cluster operation while guaranteeing the performance requirements of user requests. To do this, we first categorize the DL jobs into two classes: training jobs and inference jobs. Through the architecture-agnostic modeling, our CE-DLA approach is able to conduct the delicate mapping of heterogeneous DL jobs to GPU computing nodes. Second, we design the electricity price-aware DL job allocation so as to minimize the energy consumption cost of the cluster. We show that our approach efficiently avoids the peak-rate time slots of the GPU computing nodes by using the sophisticated mixed-integer nonlinear problem (MINLP) formulation. We additionally integrate the dynamic right-sizing (DRS) method with our CE-DLA approach, so as to minimize the energy consumption of idle nodes having no running job. In order to investigate the realistic behavior of our approach, we measure the actual output from the NVIDIA-based GPU devices with well-known deep neural network (DNN) models. Given the real trace data of the electricity price, we show that the CE-DLA approach outperforms the competitors in views of both the energy consumption cost and the performance for DL job processing.