Cloud computing environments offer on-demand access to a shared pool of configurable computing resources. Efficient resource allocation is crucial for maximizing performance, minimizing costs, and ensuring quality of service. This paper proposes a novel hybrid metaheuristic approach for optimizing resource allocation in cloud environments. The approach combines the strengths of the Genetic Algorithm (GA) and Simulated Annealing (SA) to achieve a superior balance between exploration and exploitation of the search space. The GA is used for global exploration, identifying promising regions of the solution space, while SA is employed for local refinement, fine-tuning solutions within those regions. The proposed hybrid algorithm is evaluated through simulations on a cloud environment, and the results demonstrate its effectiveness in minimizing resource utilization, reducing makespan, and improving overall system performance compared to traditional GA and SA algorithms. The results highlight the potential of the hybrid approach for practical applications in cloud resource management.
This research investigates the application of hybrid deep learning models coupled with sensor fusion techniques to enhance predictive maintenance strategies for industrial machinery. The study addresses the limitations of traditional maintenance approaches and explores the potential of advanced machine learning algorithms to accurately predict equipment failures and optimize maintenance schedules. A novel hybrid model combining Convolutional Neural Networks (CNNs) for feature extraction from time-series sensor data and Recurrent Neural Networks (RNNs), specifically Long Short-Term Memory (LSTM) networks, for temporal dependency modeling is proposed. The model is trained and validated on a comprehensive dataset obtained from various sensors monitoring the operational parameters of industrial equipment. The results demonstrate significant improvements in prediction accuracy, reduced false alarm rates, and optimized maintenance scheduling compared to conventional methods and standalone deep learning models. The proposed approach offers a promising solution for enhancing the reliability and efficiency of industrial operations through proactive maintenance strategies.
The integration of renewable energy sources (RES) into the smart grid presents significant challenges and opportunities. Intermittency and variability in RES generation, coupled with fluctuating demand, can strain grid stability and reliability. This paper proposes a hybrid forecasting model that combines machine learning techniques with statistical methods to predict renewable energy generation and dynamic load balancing strategies to enhance smart grid resilience. The forecasting model integrates Long Short-Term Memory (LSTM) networks with Autoregressive Integrated Moving Average (ARIMA) models to improve prediction accuracy. The dynamic load balancing strategy employs a multi-objective optimization algorithm, considering both cost minimization and grid stability. Simulation results demonstrate the effectiveness of the proposed approach in mitigating the impact of RES intermittency, reducing overall energy costs, and improving grid reliability under various operational scenarios. The paper concludes with a discussion of the limitations and potential future research directions in this critical area of smart grid management.
The Internet of Things (IoT) is rapidly expanding, creating numerous opportunities but also exposing critical vulnerabilities. Intrusion Detection Systems (IDSs) are crucial for securing IoT networks, yet their performance is often hampered by the high dimensionality of data generated by IoT devices. This paper proposes an adaptive hybrid metaheuristic optimization algorithm for enhanced feature selection in high-dimensional IoT intrusion detection systems. The algorithm combines the strengths of Genetic Algorithm (GA) and Particle Swarm Optimization (PSO) with an adaptive parameter control mechanism to efficiently explore the feature space and identify the most relevant features for accurate intrusion detection. The adaptive parameter control dynamically adjusts the parameters of GA and PSO based on the search progress, preventing premature convergence and improving the overall search efficiency. The proposed approach is evaluated using benchmark IoT intrusion detection datasets and compared with state-of-the-art feature selection methods. The experimental results demonstrate that the proposed algorithm achieves superior performance in terms of detection accuracy, false positive rate, and computational efficiency, making it a promising solution for securing IoT networks against evolving cyber threats.
This research investigates the optimal design and operation of an intelligent hybrid renewable energy system (HRES) for microgrid applications, focusing on enhancing both resilience and economic viability. A multi-objective optimization framework is developed, integrating HOMER Pro for initial system sizing and simulation with a Genetic Algorithm (GA) for refined optimization. The study considers solar photovoltaic (PV) panels, wind turbines, battery energy storage systems (BESS), and diesel generators as key components of the HRES. The objectives are to minimize the Levelized Cost of Energy (LCOE) and maximize system resilience, quantified by metrics such as Loss of Power Supply Probability (LPSP) and System Average Interruption Duration Index (SAIDI). The proposed methodology is applied to a case study representing a remote community in India. Results demonstrate that the optimized HRES configuration achieves a significant reduction in LCOE while simultaneously improving microgrid resilience compared to traditional grid-connected or diesel-only systems. The study highlights the importance of integrating intelligent optimization techniques for designing sustainable and reliable energy solutions for off-grid and grid-connected applications.