Energy Resources Research Articles

Tabu search based on novel neighborhood structures for solving job shop scheduling problem integrating finite transportation resources

As advancements in transportation equipment intelligence continue, the job shop scheduling problem integrating finite transportation resources (JSPIFTR) has attracted considerable attention. Within the domain of shop scheduling, the neighborhood structure serves as a cornerstone for enabling intelligent optimization algorithms to effectively navigate and discover optimal solutions. However, current algorithms for JSPIFTR rely on generalized neighborhood structures, which incorporate operators like insertion and swap. While these structures are tailored to the encoding vectors, their utilization often leads to suboptimal optimization efficacy. To address this limitation, this paper introduces novel neighborhood structures specifically designed to the distinctive properties of JSPIFTR. These innovative structures leverage the intrinsic structural information in integrated scheduling, thereby enhancing the optimization effectiveness of the algorithm. Firstly, two theorems are presented to demonstrate the feasibility of the neighborhood solution. Secondly, different neighborhood structures for critical transportation and processing tasks are subsequently designed based on the analysis of the problem properties and constraints. Thirdly, an efficient fast evaluation method is developed to expediently calculate the objective value of the neighborhood solution. Finally, the novel neighborhood structures are combined with the tabu search (TS_NNS) and compared with other state-of-the-art methods on EX and NEX benchmarks. The comparative results demonstrate the remarkable performance of the neighborhood structure, with the TS_NNS enhancing the best solutions across 23 instances.

What can European security architecture look like in the wake of Russia’s war on Ukraine?

ABSTRACT The paper outlines the most likely scenarios of the future European security architecture based on an analysis of several drivers, such as Russia’s leadership’s perception of Ukraine, the mindset of the Putinist elite, Russia’s state geopolitics, the domestic situation and prospects of a regime change in Russia, economic sanctions against Russia, energy and the export of Russia’s energy resources, and the positions of key Euratlantic actors. The current “unstable balance model” will remain the most likely scenario in case of a conservation of the situation on the Ukrainian battlefields. Since the concert model of security architecture entails a lack of a revisionist agenda among all the actors, the only reasonably probable security management upgrade is a “positive stable balance”. This scenario is conditioned on degrading Russia’s military capacity, Ukraine’s military advancements on the battlefields, a weakening of the Russian economy and maintaining the support for Ukraine in the EU and NATO.

Innovative MOS-based fiber cement boards: Effect of kraft pulp mills waste and curing by accelerated carbonation

The increasing global waste generated by industrial activities poses significant environmental challenges. Eco-waste management emerges as an economically viable solution for converting, valorizing, and repurposing these byproducts, aligning with circular economy principles, and aiming to reduce greenhouse gas emissions. Construction, known for its unsustainability due to high energy consumption, non-renewable resource utilization, waste generation, and greenhouse gas emissions, particularly from cement production, has led to the exploration of alternative materials. Magnesium oxide (MgO)-based cement, an alternative to Ordinary Portland Cement (OPC), has gained attention, leveraging Brazil's prominence as a major magnesite producer. This study explores the valorization of waste from kraft pulp mills of the paper industry, specifically lime sludge (LS) and lime slaker grits (grits), to produce magnesium oxysulfate (MOS)-based fiber cement boards. The effect of accelerated carbonation on cementitious composites produced with formulations containing grits and LS was investigated. Replacing 25% of the limestone with grits showed no noticeable differences in the properties of the boards, and MOR values close to 11.17 MPa were obtained. However, the physical-mechanical performance showed a decrease with the use of LS and higher concentrations of grits, associated with Na2SO4.XH2O formation in the system and ITZ structure formed around the aggregated particle, respectively. Carbonation in a saturated atmosphere led to the carbonation of the 5–1–7 phase, which was related to the decrease in mechanical strength of the boards after the curing process. The thermal decomposition of the Hydrated Magnesium Carbonates (HMCs) formed during accelerated carbonation corroborated with the changes in the physical properties of the composites, demonstrating that the carbonation products are formed within the voids and pores of the material and contribute to the reduction in water absorption of the boards.

Secondary flexibility market mechanism design and response behavior analysis among multi-microgrids with high proportional BTM-RERs

As the proportion of behind-the-meter (BTM) renewable energy resources (RERs) in the distribution network continues to rise, the precise role of RERs in the flexibility market, whether as “Recipients” or “Providers”, is becoming less distinct. This observation has led us to explore the integration of flexible load resources and RERs into flexibility markets. Such integration facilitates resource sharing among multiple microgrids (MGs). This paper focuses on the potential emergence of a secondary flexibility (SecFlex) market in the power sector among high proportional BTM-RERs microgrids. Within the concept of SecFlex market, this paper outlines a comprehensive process supporting secondary flexibility trading, which includes “SecFlex market organization–trading volume collation–response schedule optimization”. At first, the market mechanism of the SecFlex market is proposed, where players submit an empirical load curve to the market organizer. During the clearing stage, a social welfare equilibrium problem is solved. Subsequently, the equilibrium trading volume is adjusted in the collation stage to account for the distribution network (DN) voltage magnitude balance. Finally, a response schedule strategy is designed to analyze the behaviors of the players. The cvxpy framework is used to solve social welfare and competitive equilibrium in this paper. The results show significant cost reductions for two players, by 77.8% and 24.0% respectively. However, costs for Player 3 have increased substantially, up to 4.6 times the baseline when flexibility is limited. These findings provide theoretical guidance for the development of the SecFlex market.

Maintenance Pulmonary Rehabilitation: An Update and Future Directions.

The Global Initiative for Chronic Obstructive Lung Disease guidelines recommend pulmonary rehabilitation (PR) for individuals with COPD to improve exercise capacity and health-related quality of life (HRQOL) and reduce symptoms of dyspnea. For cost-effectiveness in COPD care, PR is second only to smoking cessation. However, PR programs typically last 9-12 weeks. The benefits of PR in terms of exercise capacity and HRQOL often decrease toward pre-PR levels as early as 3-6 months after completing PR if patients do not continue to engage in exercise. This review will (1) briefly summarize the efficacy data that informed the 2023 American Thoracic Society (ATS) clinical practice guidelines for maintenance PR, (2) discuss exercise components of maintenance PR studied since 2020 when the last papers were included in the ATS guidelines, (3) explore future directions for delivery of maintenance PR using technology-mediated models, and (4) examine the need for behavior change techniques informed by theoretical models that underpin long-term behavior change. This review will focus on persons with COPD who have completed an out-patient core initial PR program as most of the data on maintenance PR have been published in this patient population. Core PR typically implies a facility-based initial intensive structured program. All patients who complete a core initial PR program should be counseled by PR staff at the discharge visit to engage in ongoing exercise. This usual care is equally as important as referral to a formal PR maintenance program. It is critical to emphasize that usual care after core initial PR means all patients should be supported to participate in regular ongoing exercise, regardless of whether supervised maintenance PR is available. Currently, the optimal frequency, exercise and/or physical activity content, and delivery mode for maintenance PR in persons with COPD and other chronic respiratory diseases remain unknown. Patient safety and degree of in-person supervision required due to the severity of the underlying lung disease need to be considered. Future research of maintenance PR should be underpinned by behavior change techniques. Finally, in the setting of finite resources, balancing the competing priorities of core initial programs with those of maintenance PR programs needs to be achieved.

A DRL-Based Satellite Service Allocation Method in LEO Satellite Networks

Satellite computing represents a recent computational paradigm in the development of low Earth orbit (LEO) satellites. It aims to augment the capabilities of LEO satellites beyond their current transparent relay functions by enabling real-time processing, thereby providing low-latency computational services to end users. In LEO constellations, a significant deployment of computationally capable satellites is orchestrated to offer enhanced computational resources. Challenges arise in the optimal allocation of terminal services to the most suitable satellite due to overlapping coverage among neighboring satellites, compounded by constraints on satellite energy and computational resources. The satellite service allocation (SSA) problem is recognized as NP-hard, yet assessing allocation methods through results allows for the application of deep reinforcement learning (DRL) to obtain improved solutions, partially addressing the SSA challenge. In this paper, we introduce a satellite computing capability model to quantify satellite computational resources. A DRL model is proposed to address service demands, computational resources, and resolve service allocation conflicts, strategically placing each service on appropriate servers. Through simulation experiments, numerical results demonstrate the superiority of our proposed method over baseline approaches in service allocation and satellite resource utilization, showcasing advancements in this field.

Sustainable Renewable Energy Resources in Comoros Islands

Sustainable Renewable Energy Resources in Comoros Islands

Non‐linear nexus of mineral rents, coal rents, foreign direct investment, and environmental sustainability: Importance of institutional quality in E‐7 nations

AbstractThis study investigates the impact of natural resources (NRs), economic growth (GDP), institutional quality, mining and coal rents, and foreign direct investment (FDI) on the environmental quality in the E‐7 countries (Brazil, China, India, Indonesia, Russia, Turkey, and Mexico) from 1990 to 2021. To do this, such advanced estimation methods are used as the method of moments quantile regression (MMQR), with the artificial neural network utilized for robustness checks. The initial diagnostic test results unveiled the presence of cross‐sectional dependence, slope heterogeneity, and long‐run cointegration. The MMQR analysis results indicate that coal and mineral rents escalate environmental pollution by reducing the load capacity factor (LCF) across all quantiles. The results further indicate that FDI and economic growth negatively affect while institutional quality (INS) significantly promotes the LCF in E‐7 countries. Moreover, INS significantly moderates the relationship between mineral resources and LCF, as well as coal resources and LCF. The robustness test also validates these findings. Based on these results, policy measures aimed at strengthening institutional quality, particularly within the NR sector, are crucial for mitigating the adverse impacts of coal and mineral extraction on LCF and promoting sustainable development in the E‐7 countries.

Driving innovation in energy and telecommunications: next-generation energy storage and 5G technology for enhanced connectivity and energy solutions

Driving innovation in energy and telecommunications involves leveraging next-generation energy storage and 5G technology to enhance connectivity and energy solutions. This review explores the intersection of these two domains, highlighting the importance of advancements in energy storage and 5G technology for a sustainable and connected future. Energy storage is crucial for balancing the supply and demand of electricity in modern power systems. Traditional energy storage methods, such as batteries and pumped hydro, have limitations in terms of scalability, efficiency, and cost-effectiveness. Next-generation energy storage technologies, including advanced batteries, hydrogen storage, and thermal storage, offer promising solutions to overcome these limitations. These technologies enable efficient energy storage at scale, facilitating the integration of renewable energy sources like solar and wind into the grid. By storing excess energy generated during periods of low demand, next-generation energy storage systems ensure a reliable and stable power supply, reducing the reliance on fossil fuels and lowering greenhouse gas emissions. In parallel, the evolution of telecommunications technology, particularly the advent of 5G networks, is revolutionizing connectivity and communication. 5G technology offers significantly higher data transfer speeds, lower latency, and increased network capacity compared to its predecessors. These capabilities are essential for supporting emerging technologies such as the Internet of Things (IoT), autonomous vehicles, and smart grids. With 5G-enabled IoT devices, utilities can monitor energy consumption in real-time, optimize grid operations, and detect and respond to faults more efficiently. Moreover, 5G connectivity enhances the efficiency and reliability of energy storage systems by enabling seamless communication between distributed energy resources and grid operators. The convergence of next-generation energy storage and 5G technology presents numerous opportunities for driving innovation in both energy and telecommunications sectors. One of the key areas of innovation is the development of smart energy storage systems equipped with 5G connectivity. These systems can autonomously adjust their operation based on grid conditions, weather forecasts, and energy demand patterns, optimizing energy storage and distribution in real-time. Furthermore, advanced energy management algorithms leveraging artificial intelligence (AI) and machine learning (ML) algorithms can optimize energy usage and storage, further improving the efficiency and reliability of the grid. Another area of innovation lies in the integration of renewable energy resources with 5G-enabled microgrids. Microgrids are localized energy systems that can operate independently or in conjunction with the main grid. By combining renewable energy sources with energy storage and 5G-enabled communication, microgrids can provide reliable, clean, and resilient power to remote or urban areas. These microgrids can also facilitate peer-to-peer energy trading, allowing consumers to buy and sell excess energy within their communities, fostering energy independence and sustainability. Furthermore, advancements in battery technology, such as solid-state batteries and flow batteries, are enhancing the performance and reliability of energy storage systems. Solid-state batteries offer higher energy density, faster charging rates, and improved safety compared to conventional lithium-ion batteries. Flow batteries, on the other hand, provide scalability and long-duration storage capabilities, making them suitable for grid-scale applications. Integrating these advanced battery technologies with 5G-enabled monitoring and control systems enhances the overall resilience and flexibility of the energy infrastructure. In addition to technological advancements, driving innovation in energy and telecommunications requires collaboration among various stakeholders, including policymakers, regulators, industry players, and research institutions. Policies and regulations should incentivize the deployment of next-generation energy storage and 5G infrastructure, promote interoperability standards, and ensure data privacy and security. Public-private partnerships can facilitate the investment and deployment of innovative solutions, while research and development initiatives can spur further technological advancements. Driving innovation in energy and telecommunications through next-generation energy storage and 5G technology is essential for building a sustainable, connected, and resilient future. By leveraging advanced energy storage systems, smart grids, and 5G-enabled communication networks, we can optimize energy usage, reduce carbon emissions, and enhance the reliability and efficiency of our energy infrastructure. Collaboration and investment across various sectors are key to unlocking the full potential of these transformative technologies and achieving a brighter, more sustainable future for generations to come. Keywords: Innovation, Energy, Telecommunications, Next-Generation, 5G technology, Enhanced connectivity.

Data-driven optimization for microgrid control under distributed energy resource variability

The integration of renewable energy resources into the smart grids improves the system resilience, provide sustainable demand-generation balance, and produces clean electricity with minimal leakage currents. However, the renewable sources are intermittent in nature. Therefore, it is necessary to develop scheduling strategy to optimise hybrid PV-wind-controllable distributed generator based Microgrids in grid-connected and stand-alone modes of operation. In this manuscript, a priority-based cost optimization function is developed to show the relative significance of one cost component over another for the optimal operation of the Microgrid. The uncertainties associated with various intermittent parameters in Microgrid have also been introduced in the proposed scheduling methodology. The objective function includes the operating cost of CDGs, the emission cost associated with CDGs, the battery cost, the cost of grid energy exchange, and the cost associated with load shedding. A penalty function is also incorporated in the cost function for violations of any constraints. Multiple scenarios are generated using Monte Carlo simulation to model uncertain parameters of Microgrid (MG). These scenarios consist of the worst as well as the best possible cases, reflecting the microgrid’s real-time operation. Furthermore, these scenarios are reduced by using a k-means clustering algorithm. The reduced procedures for uncertain parameters will be used to obtain the minimum cost of MG with the help of an optimisation algorithm. In this work, a meta-heuristic approach, grey wolf optimisation (GWO), is used to minimize the developed cost optimisation function of MG. The standard LV Microgrid CIGRE test network is used to validate the proposed methodology. Results are obtained for different cases by considering different priorities to the sub-objectives using GWO algorithm. The obtained results are compared with the results of Jaya and PSO (particle swarm optimization) algorithms to validate the efficacy of the GWO method for the proposed optimization problem.

Operation scheduling method in an off-grid considering frequency violation probability caused by short-term variation

The use of renewable energy (RE) resources in off-grid applications is a promising solution from both the economic and environmental perspectives. However, the employment of RE generation should consider the uncertainty in their generation output; in particular, short-term variations in generation output result in serious frequency violations. The development of the off-grid planning, operation, and scheduling considering frequency variation is significant. This study proposes a probabilistic model for the frequency variation caused by load variation, which includes the dynamics of rotating generators and frequency-supporting smart inverters. The proposed method calculates the maximum frequency deviation by transforming the block, considering the output limit of the governor and smart inverters. This calculation is fast because it does not require time-sequential simulation. The effectiveness of the proposed method for off-grid operational scheduling problems was demonstrated through numerical studies. The results show that the proposed method can evaluate the same results as time series simulations 1500 times faster. Additionally, frequency deviations caused by short-term variations can be avoided, and battery capacity for zero frequency deviation is estimated with a 5% error. These results have the potential to contribute significantly to the operation of the grid with a large portion of RE resources.

Features of the Contract for Supply of Energy and Other Resources through an Attached Network

Features of the Contract for Supply of Energy and Other Resources through an Attached Network

Features of the rights and obligations of the parties to the contract for the supply of energy resources through the connected network

Features of the rights and obligations of the parties to the contract for the supply of energy resources through the connected network

Resilient data-driven non-intrusive load monitoring for efficient energy management using machine learning techniques

The integration of smart homes into smart grids presents numerous challenges, particularly in managing energy consumption efficiently. Non-intrusive load management (NILM) has emerged as a viable solution for optimizing energy usage. However, as smart grids incorporate more distributed energy resources, the complexity of demand-side management and energy optimization escalates. Various techniques have been proposed to address these challenges, but the evolving grid necessitates intelligent optimization strategies. This article explores the potential of data-driven NILM (DNILM) by leveraging multiple machine learning algorithms and neural network architectures for appliance state monitoring and predicting future energy consumption. It underscores the significance of intelligent optimization techniques in enhancing prediction accuracy. The article compares several data-driven mechanisms, including decision trees, sequence-to-point models, denoising autoencoders, recurrent neural networks, long short-term memory, and gated recurrent unit models. Furthermore, the article categorizes different forms of NILM and discusses the impact of calibration and load division. A detailed comparative analysis is conducted using evaluation metrics such as root-mean-square error, mean absolute error, and accuracy for each method. The proposed DNILM approach is implemented using Python 3.10.5 on the REDD dataset, demonstrating its effectiveness in addressing the complexities of energy optimization in smart grid environments.

Optimization and performance enhancement of renewable energy microgrid energy system using pheasant bird optimization algorithm

The main aim of this research work is to design and select an optimal renewable energy resource based microgrid (MG) system for rural area electrification of India. MG system consists of diesel generator (DG), battery bank (BB), wind turbine (WT), biomass generator (BMG) and solar photovoltaic (PV) modules with seven different configurations. The optimal system is selected based on technological, economic, environmental, social and reliability parameters. For this objective, a novel algorithm called pheasant bird optimization algorithm (PBOA) is presented and compared with particle swarm optimization (PSO), genetic algorithm (GA), ant colony optimization (ACO), cuckoo search algorithm (CSA) and HOMER Pro. Results show that PV/WT/BMG/BB/DG is an optimal configuration with sizes of 1148 units of PV, 42 units of WT, 20 units of BMG, 36 units of BB and 1 unit of DG. PBOA has determined the optimal resource size for the MG system. In terms of social and reliability parameters for the PV/WT/BMG/BB/DG system, PBOA generated values of 93.58 %, for RF, 0.08754 for EGJC and 0.01 % for LPSP. For the optimal configuration, PBOA estimated NPC, COE, CE values are ₹66185779, ₹9.3 and 650901 kg/yrs respectively. PBOA reaches the minimal NPC of ₹66185779 in just 180 iterations, whereas PSO, ACO, GA, and CSA required 360, 200, 360 and 250 iterations respectively. PBOA also achieved minimal COE at a rate of 1.52, 1.11, 1.11 and 1.94 times faster than PSO, GA, CSA and ACO respectively. The contribution of PV and WT sources to the total generated electricity is approximately 50 % and 35 % respectively. PBOA exhibits faster convergence time and iteration compared to the PSO, ACO, GA and CSA algorithm. The comparison results indicate that the optimal PV/WT/BMG/BB/DG configuration select for rural areas. This research provides valuable information for policy makers and investors on the implementation of MG system in rural regions.

Analysis of the sustainability index for ecologically low-input integrated farming: A comprehensive assessment of environmental, economic, and social impact

A comprehensive sustainability index was designed in this study to evaluate the effectiveness and environmental impact of ecologically low-input integrated farming systems. By integrating diverse agricultural practices while minimizing reliance on external, non-renewable resources, this approach aims to enhance ecological harmony and sustainability. The index provides a multi-dimensional assessment, encompassing environmental, economic, and social aspects, to gauge the overall sustainability of these integrated farming practices. The procedure involved defining critical sustainability criteria across environmental, economic, and social dimensions, collecting relevant data on farming practices, and constructing a model to assess and score these practices. The index was validated through real-world applications and engaged with various stakeholders for insights and practical verification. Environmental data involves measuring soil nutrients (5 %), water usage (500 gallons/acre), and biodiversity (30 species/acre) to assess the ecological impact of farming practices. Economic data focuses on recording crop yields, input costs, and market prices to evaluate the financial viability of these practices. Social data is collected through surveys on labor practices, job satisfaction, and community impact, providing insights into the social implications and acceptability of the farming methods. The data revealed that the farming system aligned well with sustainability benchmarks, exhibiting no deviation from optimal levels in soil nitrogen content and water usage. Strong profitability and input cost efficiency were observed in the economic dimension, with a net income per acre of $400/acre and an input cost efficiency of 50 %. In the social aspect, although reasonably scored, specific labor metrics and regional/industry standards were lacking for a comprehensive evaluation. Overall, the farming system demonstrated commendable sustainability practices, with an environmental score of 7.5/10, an economic score of 8/10, and a social score of 7/10, suggesting potential for further improvements, particularly in addressing social sustainability factors.

Multi-agent cooperative optimal scheduling strategy of integrated energy system in urban area under extreme events

Because the global climate change intensifies as well as the natural disasters frequently occur, extreme events have caused serious impacts on the energy system in urban areas, and at the same time, they have brought great challenges to the supply and scheduling of urban energy systems. Therefore, in order to better integrate and manage various energy resources in urban areas, a Deep Q-Leaning Network-Quasi Upper Confidence Bound model is innovatively constructed using deep reinforcement learning technology to learn the state and behavior mapping relationship of energy system. Use deep learning to fit complex nonlinear models to optimize the entire energy system. Compare and verify the experiment with the real energy system. The improved Deep reinforcement learning algorithm is compared with Q-learning model, PDWoLF PHC algorithm model, Quasi Upper Confidence Bound algorithm model and deep Q-Leaning Network algorithm model. The results show that the research algorithm has the smallest instantaneous error value and absolute value of frequency deviation for area control, and the average value of the research algorithm in the absolute value of the frequency deviation is reduced by 45%–73% compared to other algorithms; over time, the unit output power of the research algorithm is able to flexibly track the stochastic square wave loads. Therefore, the proposed system strategies can provide feasible solutions to meet the challenges of extreme events and promote the sustainable development and safe operation of urban energy systems.

Enhancing Ethiopian power distribution with novel hybrid renewable energy systems for sustainable reliability and cost efficiency

Economic development relies on access to electrical energy, which is crucial for society’s growth. However, power shortages are challenging due to non-renewable energy depletion, unregulated use, and a lack of new energy sources. Ethiopia’s Debre Markos distribution network experiences over 800 h of power outages annually, causing financial losses and resource waste on diesel generators (DGs) for backup use. To tackle these concerns, the present study suggests a hybrid power generation system, which combines solar and biogas resources, and integrates Superconducting Magnetic Energy Storage (SMES) and Pumped Hydro Energy Storage (PHES) technologies into the system. The study also thoroughly analyzes the current and anticipated demand connected to the distribution network using a backward/forward sweep load flow analysis method. The results indicate that the total power loss has reached its absolute maximum, and the voltage profiles of the networks have dropped below the minimal numerical values recommended by the Institute of Electrical and Electronics Engineers (IEEE) standards (i.e., 0.95–1.025 p.u.). After reviewing the current distribution network’s operation, additional steps were taken to improve its effectiveness, using metaheuristic optimization techniques to account for various objective functions and constraints. In the results section, it is demonstrated that the whale optimization algorithm (WOA) outperforms other metaheuristic optimization techniques across three important objective functions: financial, reliability, and greenhouse gas (GHG) emissions. This comparison is based on the capability of the natural selection whale optimization algorithm (NSWOA) to achieve the best possible values for four significant metrics: Cost of Energy (COE), Net Present Cost (NPC), Loss of Power Supply Probability (LPSP), and GHG Emissions. The NSWOA achieved optimal values for these metrics, namely 0.0812 €/kWh, 3.0017 × 106 €, 0.00875, and 7.3679 × 106 kg reduced, respectively. This is attributable to their thorough economic, reliability, and environmental evaluation. Finally, the forward/backward sweep load flow analysis employed during the proposed system’s integration significantly reduced the impact of new energy resources on the distribution network. This was evident in the reduction of total power losses from 470.78 to 18.54 kW and voltage deviation from 6.95 to 0.35 p.u., as well as the voltage profile of the distribution system being swung between 1 and 1.0234 p.u., which now comply with the standards set by the IEEE. Besides, a comparison of the cost and GHG emission efficiency of the proposed hybrid system with existing (grid + DGs) and alternative (only DGs) scenarios was done. The findings showed that, among the scenarios examined, the proposed system is the most economical and produces the least amount of GHG emissions.

Healthcare professional views about a prehospital redirection pathway for stroke thrombectomy: a multiphase deductive qualitative study

BackgroundMechanical thrombectomy for stroke is highly effective but time-critical. Delays are common because many patients require transfer between local hospitals and regional centres. A two-stage prehospital redirection pathway consisting of...

Quantized non-volatile nanomagnetic domain wall synapse based autoencoder for efficient unsupervised network anomaly detection

Abstract Anomaly detection in real-time using autoencoders implemented on edge devices is exceedingly challenging due to limited hardware, energy, and computational resources. We show that these limitations can be addressed by designing an autoencoder with low-resolution non-volatile memory-based synapses and employing an effective quantized neural network learning algorithm. We further propose nanoscale ferromagnetic racetracks with engineered notches hosting magnetic domain walls (DW) as exemplary non-volatile memory based autoencoder synapses, where limited state (5-state) synaptic weights are manipulated by spin orbit torque (SOT) current pulses to write different magnetoresistance states. The performance of anomaly detection of the proposed autoencoder model is evaluated on the NSL-KDD dataset. Limited resolution and DW device stochasticity aware training of the autoencoder is performed, which yields comparable anomaly detection performance to the autoencoder having floating-point precision weights. While the limited number of quantized states and the inherent stochastic nature of DW synaptic weights in nanoscale devices are typically known to negatively impact the performance, our hardware-aware training algorithm is shown to leverage these imperfect device characteristics to generate an improvement in anomaly detection accuracy (90.98%) compared to accuracy obtained with floating-point synaptic weights that are extremely memory intensive. Furthermore, our DW-based approach demonstrates a remarkable reduction of at least three orders of magnitude in weight updates during training compared to the floating-point approach, implying significant reduction in operation energy for our method. This work could stimulate the development of extremely energy efficient non-volatile multi-state synapse-based processors that can perform real-time training and inference on the edge with unsupervised data.