Network Units Research Articles

Physical information-guided multidirectional gated recurrent unit network fusing attention to solve the Black-Scholes equation

Reasonable option pricing is crucial in the financial derivatives market. Finding analytical solutions for the Black-Scholes (BS) equation, particularly for American options or with fluctuating volatility and interest rates, is challenging. BS equations exhibit strong time-series characteristics, with asset prices typically adhering to geometric Brownian motion. To address the BS equations, we propose a sequence-to-sequence model guided by physical information (PI), called PiMGA. The PiMGA fuses a multidirectional gated recurrent unit (GRU) network with an attention module, where multidirectional GRU enhances the coding performance of the input sequences and the attention module balances the feature weights of the hidden variables. Prior physical knowledge in BS equations is jointly used as a constraint, forming the penalty function for objective optimization. This allows PiMGA to serve as an efficient approximation function in the learning paradigm of physically informed machine learning to solve BS equations. BS equations with various complexities illustrate the accuracy and feasibility of PiMGA for numerical solutions. Furthermore, the out-of-distribution generalization ability of PiMGA is verified by predicting the Nasdaq 100 index.

Women attending the sexual assault treatment unit services in the Republic of Ireland: A 7-year review.

Sexual assault is pervasive in today's society, with the numbers of those reporting it increasing. In Ireland, 50% of women will experience some form of sexual violence in their lifetime. We sought to describe the incident details of females presenting to the Sexual Assault Treatment Unit (SATU) network in the Republic of Ireland and to determine associations between incident characteristics and: (1) victim age, (2) presence of injury, (3) victim-perpetrator relationship, and (4) number of assailants. This was a retrospective cross-sectional study of all females who attended between 2017 and 2023. Descriptive bivariate analysis was performed. There were 5942 female attendances, with an average age of 26 years. The largest age group was women between 18 and 25 years (38.1%, n = 2263). Forensic examinations were performed in 76.6% (n = 4549). Assailants were male in 92% (n = 5469) of incidents, with multiple assailants disclosed in 7.3% (n = 435). Strangers or recent acquaintances were the assailant in 38.5% (n = 2290) of incidents, and close associates in 22.9% (n = 1359). Incidents occurred at the survivor's home in 22% (n = 1306) of incidents, and in the assailant's home in 22.6% (n = 1342). Drug use within 24 h was reported in 15.1% (n = 897) of cases, and alcohol use in 72% (n = 4276). Drug-facilitated assault was suspected by 16.1% (n = 955). Injuries (genital and extra-genital) occurred in 30.3% (n = 1800) of attendances and were more likely to be seen in those who disclosed ingesting alcohol (Relative risk [RR] 1.325, P < 0.001) or drugs (RR 1.111, P = 0.04) in the 24 h preceding the incident, in those who presented within 24 h (RR 1.646, P < 0.001), those aged 18 years or older (RR 1.07, P = 0.003), and those where the incident occurred outdoors (RR 1.24, P < 0.001). This study, one of the largest on female attendances to a national SATU network, offers detailed insights into demographics, incident details, and circumstances. Most were young women, mainly 18 years and older, many of whom were full-time students. Forensic examinations were the primary reason for attendance, underscoring the network's key role in evidence collection. The study also identified factors linked to a higher risk of injury detection, such as immediate post-incident attendance, being over 18, outdoor incidents, perpetration by a stranger, and prior alcohol or drug use.

Coherent time- and wavelength-division multiplexed point-to-multipoint optical access network using low-cost DFB lasers enabled by a frequency comb

Coherent reception, along with time- and wavelength-division multiplexing (TWDM), is a promising concept to simultaneously support multiple services in future high-speed point-to-multipoint passive optical networks (PONs). The next-generation PON 2 (NG-PON2) standard describes a TWDM-PON based on IM/DD intensity modulation and direct detection (IM/DD) which employs tunable-lasers and optical filters such as tunable optical filters or cyclic arrayed-waveguide gratings. Here, we investigate what we believe to be a novel coherent TWDM-PON architecture based on a frequency comb source in the optical line terminal (OLT), and thermally-tuned distributed-feedback (DFB) lasers in the optical network units (ONUs). For downstream operation, we broadcast multiple copies of two 25 GBd dual-polarization quadrature phase shift keying (DP-QPSK) signals, resulting in a total PON downstream capacity of 200 Gbit/s. The copies of the downstream signals are spanning ±2 nm (±250 GHz). In the ONUs, we align the wavelengths of the DFB lasers, which act as local oscillators (LOs), to one of the downstream signals by a thermal heater or by changing the direct current. In upstream, the already aligned DFB lasers act as transmit lasers and the frequency comb as LO. We demonstrate TWDM upstream by emulating two ONUs with 25 GBd DP-QPSK, resulting in a total PON upstream capacity of 200 Gbit/s.

Leveraging advanced AI algorithms with transformer-infused recurrent neural networks to optimize solar irradiance forecasting

Solar energy (SE) is vital for renewable energy generation, but its natural fluctuations present difficulties in maintaining grid stability and planning. Accurate forecasting of solar irradiance (SI) is essential to address these challenges. The current research presents an innovative forecasting approach named as Transformer-Infused Recurrent Neural Network (TIR) model. This model integrates a Bi-Directional Long Short-Term Memory (BiLSTM) network for encoding and a Gated Recurrent Unit (GRU) network for decoding, incorporating attention mechanisms and positional encoding. This model is proposed to enhance SI forecasting accuracy by effectively utilizing meteorological weather data, handling overfitting, and managing data outliers and data complexity. To evaluate the model’s performance, a comprehensive comparative analysis is conducted, involving five algorithms: Artificial Neural Network (ANN), BiLSTM, GRU, hybrid BiLSTM-GRU, and Transformer models. The findings indicate that employing the TIR model leads to superior accuracy in the analyzed area, achieving R2 value of 0.9983, RMSE of 0.0140, and MAE of 0.0092. This performance surpasses those of the alternative models studied. The integration of BiLSTM and GRU algorithms with the attention mechanism and positional encoding has been optimized to enhance the forecasting of SI. This approach mitigates computational dependencies and minimizes the error terms within the model.

An innovative attention infused- BiRecurrenTwin network assisted hybrid segmentation technique for accurate heart disease prediction

Heart disease is a noteworthy global health concern, profoundly affecting public health and individual well-being. Early prediction and diagnosis are crucial for reducing its negative impact. This work focuses on developing an effectual heart disease detection employing advanced optimization approaches and data processing. The proposed system incorporates a hybrid segmentation algorithm, Hybrid K-means-Fuzzy-C-means (HKMFCM) clustering, which assigns membership degrees to data points, enabling soft clustering and accommodating data uncertainty. Additionally, the Artificial Bee Colony (ABC) Optimization method is applied for feature selection, mimicking the foraging behavior of honeybees to identify the most discriminative attributes from the dataset. This optimization algorithm iteratively explores the feature space to select features that enhance the model's predictive accuracy. Furthermore, a novel classification architecture, termed the Attention-infused BiRecurrenTwin Network, is introduced to accurately predict heart disease based on segmented and extracted patient data profiles. This classifier leverages both Bidirectional Gated Recurrent Unit (BiGRU) and Bidirectional Long Short-Term Memory (BiLSTM) networks, with their bidirectional nature capturing both past and future contexts, thus increasing the classifier's capability to detect subtle temporal patterns in patient data. In addition, the proposed system addresses traditional approaches limitations through its advanced components like HKMFCM (effectively manages data uncertainty by enabling soft clustering), ABC Optimization technique (enabling an iterative, global search for the most discriminative attributes) Attention-infused BiRecurrenTwin Network (surpasses conventional classifiers by capturing both past and future temporal patterns in patient data). The simulation outcomes demonstrate that the developed system attains improved performance, with accuracy, ROC, F-measure, precision, recall, sensitivity, and specificity values of 96.09%, 97%, 95%, 96.6%, 94.3%, 95.14%, and 97.05%, respectively which is higher than the baseline models with an average of 7.75 % increase in accuracy. These results indicate that the developed predictive models show promise in accurately classifying individuals based on their extracted features, thereby facilitating the early recognition of heart disease.

Influence of supply temperature and booster technology on the energetic performance and levelized cost of heat of a district heating network with central heat pump

The development of new low-temperature district heating networks in existing urban areas can accelerate the replacement of fossil-fuel-based heating systems by collective heating using electricity or renewable heat sources. Buildings can either be connected to the network in their current, non-refurbished state for rapid deployment or be refurbished before connection to achieve maximum energy efficiency. To assess the difference between these two strategies, this paper examines eight low-temperature district heating concepts with central heat pump, comparing their energy use and levelized cost of heat. The study includes four scenarios for each strategy, considering supply temperatures ranging from 10 °C to 75 °C. Heat pumps or electric heaters are used as booster technology when the network supply temperature is too low for space heating or domestic hot water purposes. The use of booster heat pumps shows the highest seasonal coefficient of performance for both refurbished (4.61) and non-refurbished buildings (3.43). However, booster technologies result in the highest levelized cost of heat, driven mainly by the high investment cost of the network and booster units. Lowest levelized cost of heat of 213 €/MWhth is obtained for non-refurbished buildings at 75 °C and 297 €/MWhth for refurbished buildings at 55 °C without boosters.

ADVANCED FRAUD DETECTION IN CARD-BASED FINANCIAL SYSTEMS USING A BIDIRECTIONAL LSTM-GRU ENSEMBLE MODEL

This article addresses the challenges of fraud in card-based financial systems and proposes effective detection and prevention strategies. By leveraging recent data analytics and real-time monitoring, the study aims to enhance transaction security and integrity. The authors review existing fraud detection methodologies, emerging trends, and the evolving tactics of fraudsters, emphasizing the importance of collaboration among financial institutions, regulatory agencies, and technology providers. Our proposed solution is an ensemble model combining Bidirectional Gated Recurrent Unit (BiGRU) and Bidirectional Long Short-Term Memory (BiLSTM) networks, designed to capture complex transactional patterns more effectively. Comparative analysis of six machine learning classifiers—AdaBoost, Naïve Bayes, Decision Tree, Logistic Regression, Random Forest, and Voting—demonstrates that our BiLSTM-BiGRU ensemble model outperforms traditional methods, achieving a fraud detection performance score of 89.22%. This highlights the advanced deep learning model's superior ability to enhance the robustness and reliability of fraud detection systems.

The Comparison of GRU and LSTM in Solar Power Generation Forecasting Application

Precise forecasting of solar power generation is essential for optimizing the incorporation of renewable energy into the electrical grid and maintaining the stability of energy systems. This paper offers a thorough comparative examination of Long Short-Term Memory (LSTM) networks and Gated Recurrent Unit (GRU) networks in forecasting solar power generation. Both models were assessed for predicted accuracy and computational efficiency, employing essential performance indicators including Mean Absolute Error (MAE) and R-squared (R²). The efficacy of LSTM and GRU in forecasting is assessed using a real-world dataset. Comprehensive experimental data demonstrate that the GRU model surpasses the LSTM model.

Cryptocurrencies Price Prediction Using Deep Learning Models (Gated Recurrent Unit And Recurrent Neural Network))

Background: With their volatile prices, cryptocurrencies have become valuable assets in the financial market. Predicting cryptocurrency prices accurately is essential for making well-informed investment decisions. Time series prediction models, like Gated Recurrent Unit (GRU) and Recurrent Neural Networks (RNN), are popular tools for financial data forecasting because they can capture sequential dependencies in data. Aim: This study aims to predict the average monthly closing prices of five major cryptocurrencies—Bitcoin (BTC), Ethereum (ETH), Binance Coin (BNB), Litecoin (LTC), and Ripple (XRP)—using GRU and RNN models and evaluate their performance in forecasting these prices. Method: Time series input sequences were produced and historical price data for the chosen cryptocurrencies were preprocessed using Min-Max Scaling. This data was divided into training and test sets, and it was used to train both the GRU and RNN models. The Mean Absolute Percentage Error (MAPE) and Root Mean Squared Error (RMSE) were used to assess the performance of the model. Results: For the majority of cryptocurrencies, the RNN model exhibited better predicted accuracy and consistently outperformed the GRU model. For instance, the RMSE for Ripple was 0.06 for the RNN model and 0.09 for GRU. In a similar vein, the RNN model outperformed the GRU model with a MAPE of 12.97% for Ethereum. These results imply that RNN models are more suitable for financial forecasting in this sector, as they yield more accurate predictions for cryptocurrency values.

Deep Reinforcement Learning in Cryptocurrency Trading: A Profitable Approach

This study proposes an Automatic Cryptocurrency Trading System using Deep Reinforcement Learning (DRL). Six popular cryptocurrencies were used: Bitcoin, Ethereum, BinanceCoin, DogeCoin, Cardano, and WAVES. Development of the trading system started with building three timeseries models – Temporal Convolutional Neural Network (TCNN), Long Short-Term Memory Network (LSTM), and Gated Recurrent Unit Network (GRU) – to predict future prices. Then, cryptocurrency sentiment data was scraped using the Alternative.me API. Data on historical prices, predicted future prices, cryptocurrency sentiment index, technical indicators, and trading account information was fed as input states to three DRL Agents — Deep Q Network (DQN), Advantage Actor Critic (A2C), and Recurrent Proximal Policy Optimization (RPPO) — which were trained using a custom-developed trading environment. Each agent was given $1000 initial capital for all six cryptocurrencies to trade using three possible actions — Buy, Sell and Hold — and were back-tested on one year of unseen data. Our DQN model had the highest overall return on investment (ROI) of $740, an average 12.3% ROI across all six cryptocurrencies, with an ROI of 63.98% achieved for BinanceCoin. However, A2C and RPPO both had negative ROI.

Densification mechanism and structural transformation in glass: molecular dynamic simulation

The structure of GeO2 glass is investigated via molecular dynamic simulation. The influence of pressure to the structure of GeO2 glass is due to the change of short-range order, network structure, domain structure and free volume regions. The obtained results show that the structure of GeO2 was formed by a continuous random network of basic structural units linking to each other via with corner-sharing, edge-sharing, face-sharing bond. At low pressure, the basic structural units are mainly linked via one bridge oxygen (the corner sharing bonds). As pressure increases, the fraction of edge sharing bonds (two bridge oxygen) increases meanwhile the fraction of corner sharing bonds significantly decrease. There is a slight increase with the fraction of face- sharing bonds (three bridge oxygen). The Ge-O bond distance is almost not dependent on pressure, meanwhile the Ge-Ge and O-O bond distances are significantly dependent on pressure. Under compression, the average coordination number of Ge tends to increase from fourfold (at ambient pressure) to six-fold (at high pressure). Moreover, structure of GeO2 glass has two-domain near 15 and 25 GPa. And at ambient pressure or high pressures, the struture of GeO2 glass has single domain. The dependence of the distribution of free volumes on pressure in GeO2 glass has been examined. The distribution of void radius has the Gaussian form and the position of the peak tend to shift to the left under compression.

Study on the planning and scheduling strategies of the hydrogen supply network for road transportation in deep decarbonization scenarios

As the energy decarbonization continues to advance, hydrogen is gaining momentum. With the increasing demand for hydrogen and the lack of hydrogen infrastructure, it is urgent to solve the problem of hydrogen supply network layout planning. In order to reduce redundant construction of hydrogen supply network infrastructure and increase the rationality of its layout, this paper conducts research on the optimal layout planning of hydrogen supply network infrastructure in the transportation sector. By constructing a two-stage real-time scheduling model of hydrogen supply network, the unit cost composition of hydrogen supply network is more complete and accurate, and the location, technology type and quantity of each link of hydrogen supply network are obtained. The results show that the unit cost of China's hydrogen supply network is 27.8 CNY/kg in 2035 and 19.04 CNY/kg in 2050 under the baseline scenario. Compared with the results of existing literature, the two-stage real-time scheduling model can reduce by 7 % of the hydrogen energy supply network unit cost by decreasing the transportation vehicles and the investment capacity of fixed hydrogen storage equipment. The proposed method of the hydrogen supply network planning can provide policy-makers with the selection of sustainability pathways for dynamic hydrogen infrastructure development planning.

Structural Expansion and Enhanced Photocurrent Conversion of Selenido Stannates with Cu+ Ions.

As a means of tuning the electronic properties of tin-chalcogenide-based compounds, we present a strategy for the compositional and structural expansion of selenido stannate frameworks under mild conditions by introducing Cu+ ions into binary anionic Sn/Se aggregates in ionothermal reactions. The variable coordination modes of Cu+-contrasting with tetrahedral {SnSe4} or trigonal bipyramidal {SnSe5} units-and corresponding expansion toward ternary Cu/Sn/Se substructures helped to add another degree of freedom to the nanoarchitectures. As desired, the variation of the structural features was accompanied by concomitant changes of the physical properties. Upon treatment of alkali metal salts of the [SnSe4]4- anion at slightly elevated temperatures (120 or 150 °C) in ionic liquids, we isolated a series of compounds comprising ternary or quaternary cluster molecules or networks of cluster units, (C2C2Im)9Li[Cu10Sn6Se22] (1), (C2C2Im)4[Cu8Sn6Se18] (2), (C2C1Im)3[Cu5Sn3Se10] (3), and (C2C2Im)5[Cu8Sn6Se18F]·(C2C2Im)[BF4] (4; C2C2Im = 1,3-diethyl-imidazolium, C2C1Im = 1-ethyl-3-methyl-imidazolium), which were investigated in terms of their optical gaps and photocurrent conversion properties. As illustrated by the synthesis and characterization of an additional salt that does not include Cu+, {(C2C2Im)2[Sn3Se7]}4·{(C2C2Im)[BF4]}2 (5), the significant role of Cu+ in this system was shown to be 3-fold: (a) structural expansion, (b) narrowing of the optical gap, and (c) photocurrent enhancement. By this three-in-one effect, the work offers an in-depth understanding of chalcogenido metalate chemistry with atomic precision.

GNSS elevation time series prediction model based on multiple gate recurrence unit and temporal convolutional networks parallel

GNSS elevation time series prediction model based on multiple gate recurrence unit and temporal convolutional networks parallel

Low-latency 100 Gb/s PAM-4 PON with a 42.5 dB power budget over the 20 km anti-resonant hollow-core fiber.

The surging growth in data traffic has necessitated the development of higher-speed, lower-latency intensity modulation and direct detection (IM/DD) passive optical networks (PONs) with higher power budgets. To address the inherent limitations of traditional silica-based solid-core fibers, anti-resonant hollow-core fibers have garnered significant attention from both academia and industry. In this Letter, we present an experimental demonstration of 100 Gb/s PAM-4 IM/DD PON transmission over a 20 km anti-resonant hollow-core fiber in the C band, utilizing low-complexity digital signal processing (DSP). The result achieves a record high power budget of 42.5 dB, facilitated by a 3-tap weighted lookup table (LUT) at the optical line terminal (OLT) side and a semiconductor optical amplifier used as a preamplifier at the optical network unit (ONU) side. This represents the highest power budget reported for IM/DD PON to date, to the best of our knowledge,and offers a promising alternative for the future evolution of PON systems.

Первичное звено государственной системы здравоохранения в Российской Федерации: организационно-правовой статус

Primary care is the backbone of the healthcare system. The main function of primary care is the provision of primary health care to the population on a territorial-precinct basis through “first contact” doctors. There is a gap between the importance of primary health care and low public satisfaction with it. One of the main reasons for this current state of affairs is the inadequate organizational and legal status of primary care. The declared formation of a three-tier healthcare system must be based on a more understandable architectural network of healthcare units. It is necessary to harmonize between the types of objects and the types of care they provide, the conditions of provision and forms of medical care.

Driving Forces in the Formation of Paracetamol Cocrystals and Solvate with Naphthalene, Quinoline and Acridine.

Paracetamol is an important analgesic and antipyretic drug showing poor tabletability. Among the various approaches used to improve this property, understanding the forces that govern the crystal packing is revealed to be crucial. We prepared three stable compounds: (par)2∙(nap) (1), (par)∙(quin) (2), and (par)∙(acr) (3) (nap-naphthalene, quin-quinoline, acr-acridine) being cocrystals or solvate. The structural studies showed that all the reported compounds are composed of alternately arranged layers of paracetamol and coformer. Several supramolecular motifs in the paracetamol layer were identified: R44(22) in (1); R64(20) and R22(8) in (2); and R22(8), R42(12), and R44(26) rings in (3). The stability of the crystal network was studied by interactions analysis performed by Hirshfeld surface and fingerprint approaches and the energy between the closest units in the crystal network was calculated. It showed that the strongest interactions were found between blocks connected by N-H⋯O=C and O-H⋯O/N hydrogen bonds due to an important coulombic factor. The dispersive energy becomes important for tail-to-tail (and head-to-tail) arranged paracetamol units, and it prevails in the case of stacking interactions between coformer molecules. The importance of dispersive forces increases with the size of the aromatic system of the coformer. XAS studies confirmed the successful preparation of compounds and provided some details about electron structure.

Transient NOx emission modeling of a hydrogen-diesel engine using hybrid machine learning methods

One promising approach to reduce carbon foot print of internal combustion engines (ICEs) is using alternative fuels like hydrogen, particularly by converting medium and heavy-duty diesel engines to dual-fuel hydrogen-diesel engines. To minimize elevated NOx emissions from hydrogen-fueled engine, fast and accurate emission models are essential for engine model-based control and for engine calibration and optimization using hardware-in-the-loop (HIL) setups. In this study, a fast-response NOx emissions sensor is used to measure the transient NOx emissions from a dual-fuel hydrogen-diesel engine. Subsequently, steady-state models (SSMs), quasi steady-state models (QSSMs), and transient sequential models (TSMs) in the form of black-box (BB) and gray-box (GB) models are developed for transient NOx emissions prediction. GB models utilize both information from a one dimensional (1D) physical engine model and experimental data for training, while BB models only use experimental data. SSMs are optimized artificial neural networks (ANNs) trained using steady-state data, QSSMs are optimized ANNs trained using transient data, and TSMs are time-series networks trained using transient data. Long short-term memory (LSTM) and gated recurrent unit (GRU) networks are used as the time-series deep learning networks. The results showed that the 1D physical model has the poorest performance with successive model performance improvement from SSM to QSSM and from QSSM to TSM. The developed BB TSM model in this study can predict transient NOx emissions with an R2 value greater than 0.96 at 89,000 predictions per second which makes this model suitable for real-time engine model-based control where computational efficiency is crucial. The developed GB TSM model can predict transient NOx emissions with an R2 value greater than 0.97 but it is computationally more expensive. The extra accuracy of the GB TSM models makes them the best choice for HIL setups where more computational power is available, and accuracy is more crucial.

Forecasting carbon dioxide emissions in Chongming: a novel hybrid forecasting model coupling gray correlation analysis and deep learning method.

Predicting regional carbon dioxide (CO2) emissions is essential for advancing toward global carbon neutrality. This study introduces a novel CO2 emissions prediction model tailored to the unique environmental, economic, and energy consumption of Shanghai Chongming. Utilizing an innovative hybrid approach, the study first applies grey relational analysis to evaluate the influence of economic activity, natural conditions, and energy consumption on CO2 emissions. This is followed by the implementation of a dual-channel pooled convolutional neural network (DCNN) that captures both local and global features of the data, enhanced through feature stacking. Gated recurrent unit (GRU) network then assesses the temporal aspects of these features, culminating in precise CO2 emission predictions for the region. The results indicate: (1) The proposed hybrid model achieves accurate predictions based on accounting data, with high precision, low error, and good stability. (2) The study found an overall increase in Chongming's carbon emissions from 2000 to 2022, with the prediction results being generally consistent with existing research findings. (3) The proposed method, based on Chongming's CO2 emission predictions, addresses issues such as the scarcity of effective accounting data and inaccuracies in traditional calculation methods. The results can provide effective technical support for local government policies on carbon reduction and promote sustainable development.

Digital pre-equalization for performance improvement in coherent PON

As high-speed optical access networks continue to evolve, reducing the computational complexity at the optical network unit (ONU) side remains a significant challenge in coherent passive optical network (PON). This paper presents a pre-equalization scheme that integrates digital pre-distortion (DPD) and pre-emphasis techniques in 200 Gb/s coherent PON. The focus is on three modulation formats: 50 Gbaud polarization multiplexing-quadrature phase shift keying (PM-QPSK), 50 Gbaud Alamouti-coded polarization multiplexing-16 quadrature amplitude modulation (PM-16QAM), and 25 Gbaud PM-16QAM, using intradyne or heterodyne detection schemes. Through comprehensive numerical simulations and experimental evaluations, we demonstrate significant improvements in receiver sensitivity and power budget in 200 Gb/s coherent PON. When the number of digital filtering taps is small, the proposed scheme achieves power budget enhancements of 3 dB for 50 Gbaud PM-QPSK, 7 dB for 50 Gbaud Alamouti-coded PM-16QAM, and 2.1 dB for 25 Gbaud PM-16QAM in the downlink. These findings highlight the potential of pre-equalization techniques in enhancing the performance of next-generation coherent PON with low computational complexity at the ONU side.