Faculty Publications

Permanent URI for this communityhttp://192.168.24.11:4000/handle/123456789/7

Scholarly Publications by Integral Academia

Browse

Subject: search.filters.subject.Optimization

Search Results

Now showing 1 - 3 of 3
  • Thumbnail Image
    Item
    Hybrid Particle Swarm and Gravitational Search Optimization for Intelligent Battery Health Estimation
    (IEEE, 2026) M. S. Hossain Lipu, Md Ibrahim, Shaheer Ansari, Kashem Muttaqi, Danny Sutanto
    This paper presents a novel hybrid optimization framework for data-driven estimation of the state of health (SoH) of lithium-ion batteries (LIBs). Existing data-driven SoH estimation methods struggle to select suitable model hyperparameters that consider capacity regeneration phenomena and to identify meaningful data samples from LIB parameters. The specific contribution to this research lies in the integrated optimization strategy, which bridges the limitations mentioned. The dataset has 31 features, comprising MIT-Stanford lithium-ion battery profiles. It employs an intelligent hybrid approach that combines the gravitational search algorithm (GSA) with particle swarm optimization (PSO) to fine-tune recurrent neural network (RNN) parameters, such as the hidden layer neurons and learning rate. The combined GSA and PSO algorithms integrated with RNN improve the SoH estimation accuracy through better search efficiency and faster convergence. The 31 LIB parameter samples are closely linked to capacity degradation and are well-suited to form the data framework. The proposed model demonstrates high accuracy in SoH estimation, particularly when applied to cell c33 from MIT-Stanford lithium-ion battery profiles. Results show that the RNN optimized with the GSA-PSO algorithm for the c33 dataset achieved a mean squared error (MSE) of 1.30×10−8, a root mean square error (RMSE) of approximately 0.0142, and a mean absolute percentage error (MAPE) of 0.0096.
  • Thumbnail Image
    Item
    Advances in AI and Smart Computing
    (Book Rivers, 2025) Shish Ahmad, Sheeba Praveen, Anum Kamal, Roshan Jahan, Mohammad Suaib
    This book presents recent advances in engineering, technology, and applied sciences, highlighting innovative methodologies, emerging tools, and interdisciplinary research outcomes. It brings together contributions from researchers and practitioners focusing on contemporary challenges and solutions in areas such as intelligent systems, optimization techniques, data-driven technologies, and sustainable engineering practices. The volume aims to bridge theoretical developments with real-world applications and serve as a reference for scholars and professionals seeking insights into cutting-edge advancements.
  • Thumbnail Image
    Item
    Maximizing Efficiency and Sustainability in Concrete Building Frames through Multi- Objective Simulated Annealing
    (Jupiter Publications Consortium, 2023) Tabish Izhar; Syed Aqeel Ahmad
    In the construction business, concrete building frames are a standard structural system. It is essential to optimize their design to achieve structural efficacy, cost-effectiveness, and environmental sustainability. To improve the performance of concrete building frames, this study investigates the use of Multi-objective Simulated Annealing (MOSA), a potent optimization approach. The paper thoroughly analyses MOSA, its integration with concrete frame design, and its potential advantages in getting the best results while juggling competing goals. The multiobjective optimization of reinforced concrete framed structures, which are often utilized in building construction, is the topic of this research. It demonstrates the effectiveness of a multiobjective simulated annealing (MOSA) technique used to solve two objective functions related to the financial cost of the frames and the number of bars in the reinforcement arrangement. Since bar structures with fewer bars are simpler to create, the latter aim is significant regarding the frame's constructability. The IS-456 for Structural Concrete is followed in the evaluation of solutions. Using an internal matrix technique, stress resultants and envelopes of framed structures are calculated. The national codes for building structures are followed while determining design loads.