Yazar "Ates, Abdullah" seçeneğine göre listele

Listeleniyor 1 - 2 / 2
  • Küçük Resim Yok
    Öğe
    Energy consumption model with real-time data for driving range extension of electric buses
    (Elsevier, 2025) Ekici, Yunus Emre; Aydin, Ahmet Arif; Karadag, Teoman; Akdag, Ozan; Ates, Abdullah
    Preventing range anxiety in electric vehicles (EVs) requires efficient energy use and an accurate estimation of the battery capacity needed for the desired range. A longer range leads to reduced consumption and extends operational activities. Thus, extended driving range can be achieved, promoting a more environmentally sustainable transportation model. This contributes significantly to reducing greenhouse gas emissions and mitigating the environmental impact of transportation. In this study, 250,000 rows of real-world data were collected from electric Trolleybus vehicles for a realistic energy consumption estimation of EVs. First, a mathematical model was obtained from these data using Gaussian Process Regression (GPR) method. To reduce the error rate of this model and increase the accuracy of consumption estimation, it was necessary to re-analyze it with an optimization technique. The accuracy of the consumption prediction model is extremely important for increasing the range of EVs and enabling uninterrupted travels. To solve range anxiety problem, the mathematical model obtained by GPR method is re-optimized by SeaHorse optimization and a new energy consumption prediction model, SHO-EBECM (Seahorse Optimized-Electric Bus Energy Consumption Model), is obtained. The trained SHO-EBECM was applied to 20 real routes of public transportation with internal combustion engine buses in a metropolitan city and the RMSE (Root Mean Square Error) value has been calculated to be between 0.1470 and 0.2920. Based on the achieved error rate, it can be inferred that SHO-EBECM offers a solution with a reduced error rate in comparison to four other optimization techniques. Furthermore, considering global warming, carbon emissions and ecological balance, it is concluded that approximately 12,060 tons/year of CO2, 372.75 tons/ year of NO and NO2 gases can be prevented from being emitted to nature by converting internal combustion engine buses on 20 different routes to electric buses (E-Bus) with the help of SHO-EBECM.
  • Küçük Resim Yok
    Öğe
    Optimization of distribution function and model parameters for molecular communication via diffusion with OtoO approximation
    (Elsevier, 2024) Akpamukcu, Mehmet; Ates, Abdullah; Isik, Ibrahim; Isik, Esme
    The analysis is generally conducted in stationary receiver and transmitter models in a diffusion environment for the fundamental Molecular communication (MOC) models. However, a mobile MOC model is employed in this study, deviating from the existing literature. This mobile MOC model considers the mobility of all variables in the diffusion environment, including the transmitter, receiver, and molecules. Firstly, a novel MOC model is proposed, departing from the conventional normal distribution for the mobility of variables. Instead, alternative distribution functions such as the Pareto distribution, extreme value distribution, t-distribution, and generalized extreme value distribution are employed. Furthermore, the system's performance is enhanced by optimizing the distribution function and model parameters, such as the diffusion coefficient, using the optimization of optimization (OtoO) approach. In this approach, the Multi-Verse Optimization (MVO) algorithm serves as the primary algorithm, while the Grey Wolf Optimization (GWO) algorithm functions as the auxiliary algorithm. Essentially, the MVO algorithm optimizes the parameters of the MOC model, while simultaneously, the GWO algorithm optimizes the impact of the optimization processes of MVO on the parameters ``p and ``N as well as the constant parameter of the distribution function. By optimizing both the parameters of the MOC model and the distribution function, the number of received molecules is significantly increased. Therefore, this study not only improves the results of the MOC model structure based on different distribution functions but also optimizes all parameters of the proposed model using the MVO-GWO OtoO approach.