Parallel Hybrid Architecture Research Articles

Modern approaches to data storage: comparison of relational and cloud data warehouses using etl and elt methods

The paper analyses various aspects of the use of relational and cloud data warehouses as well as methods of integrating ETL and ELT data. A comparative analysis of these approaches, their advantages and disadvantages are provided. A central relational data warehouse is proposed that provides a single version of truth (SVOT), which allows standardising and structuring data, avoiding differences and providing the access to the same information for all users of an organisation. It is analysed the methodological approaches to implementing a data warehouse: top-down, bottom-up, and from middle. It is described cloud data warehouses that use cloud technologies to provide scalability, availability and fault tolerance, which is important for the companies with huge amounts of data. The advantages and disadvantages of ETL and ELT are analysed: ETL transforms data before it is loaded into the warehouse, which makes it easier to maintain data confidentiality. ELT performs transformation after loading, which allows for more flexible data processing directly in the warehouse. In the article, we deal with the approaches to implementing a data warehouse: top-down is suitable for strategic planning, bottom-up allows for faster results, and the middle approach combines both methods to achieve optimal efficiency. We considered cloud data storage: compared to relational storage, cloud storage is more flexible, scalable and efficient, providing speed and reducing infrastructure costs. It is described cloud storage architectures: massive parallel processing, hybrid architectures, lambda architectures, and multi-structured architectures. They provide high performance and flexibility in data processing. It is described data storage technologies: Data Lake, Polyglot Persistence, Apache Iceberg, Apache Parquet, and columnar databases that provide efficient storage and processing of large amounts of data

EPHA: An Energy-efficient Parallel Hybrid Architecture for ANNs and SNNs

Artificial neural networks (ANNs) and spiking neural networks (SNNs) are two general approaches to achieve artificial intelligence (AI). The former have been widely used in academia and industry fields; the latter, SNNs, are more similar to biological neural networks and can realize ultra-low power consumption, thus have received widespread research attention. However, due to their fundamental differences in computation formula and information coding, the two methods often require different and incompatible platforms. Alongside the development of AI, a general platform that can support both ANNs and SNNs is necessary. Moreover, there are some similarities between ANNs and SNNs, which leaves room to deploy different networks on the same architecture. However, there is little related research on this topic. Accordingly, this article presents an energy-efficient, scalable, and non-Von Neumann architecture (EPHA) for ANNs and SNNs. Our study combines device-, circuit-, architecture-, and algorithm-level innovations to achieve a parallel architecture with ultra-low power consumption. We use the compensated ferrimagnet to act as both synapses and neurons to store weights and perform dot-product operations, respectively. Moreover, we propose a novel computing flow to reduce the operations across multiple crossbar arrays, which enables our design to conduct large and complex tasks. On a suite of ANN and SNN workloads, the EPHA is 1.6× more power-efficient than a state-of-the-art design, NEBULA, in the ANN mode. In the SNN mode, our design is 4 orders of magnitude more than the Loihi in power efficiency.

Hybrid convolutional long short‐term memory models for sales forecasting in retail

AbstractThis study proposes novel sales forecasting approaches that merge deep learning methods in a hybrid model. Long short‐term memory (LSTM) is adopted for modeling the temporal characteristics of the data, whereas the convolutional neural network (CNN) focuses on identifying and extracting relevant exogenous information. We propose stacked (S‐CNN‐LSTM) and parallel (P‐CNN‐LSTM) hybrid architectures to understand complex time series data with varying seasonal patterns and multiple products correlations. The performance drivers of both architectures were empirically tested with a real‐world multivariate retail dataset and outperformed when compared with simple neural network architectures and standard autoregressive methods for short and long‐term forecasting horizons. When compared with traditional predictive approaches, the proposed hybrid models reduce the computational complexity while providing flexibility and robustness.

Reducing CO2 Emissions of Hybrid Heavy-Duty Trucks and Buses: Paving the Transition to Low-Carbon Transport

This study investigates the CO2 reduction potential of powertrain hybridisation on heavy-duty lorries and city buses. The analysis considers modern parallel and serial hybrid architectures, assessing their efficiency and limits in CO2 emission reduction through vehicle simulation in VECTO, which is the official tool of the European Commission for calculating heavy-duty vehicle fuel and energy consumption. The results reveal distinct trends for each vehicle type and architecture. In lorries, more significant improvements are observed in urban delivery profiles, reaching up to ~16%, indicating the benefits of hybridisation in transient conditions with energy recuperation opportunities. City buses, particularly those with serial architectures, exhibit significant emission reductions that reach 36%, making them suitable for urban environments. The optimisation of electric motor size and performance plays a crucial role in achieving emission reductions, while battery capacity must be carefully considered to avoid adverse effects. For lorries in urban delivery use, further improvements of 17.5% can be achieved by utilising a 160 kW engine motor and 30 kWh battery. Buses are already quite well optimised, with serial architecture presenting the highest benefits with a 120 kW electric motor and a battery of 11 kWh. Future research should focus on supercapacitors and gearboxes to improve efficiency at higher vehicle speeds and assess hybridisation potential in interurban coach travel. The heavy-duty vehicle sector can make significant strides towards low-carbon transport by maximising hybrid powertrain efficiency and emission reductions.

Stochastic Gradient Descent for matrix completion: Hybrid parallelization on shared- and distributed-memory systems

The purpose of this study is to investigate the hybrid parallelization of the Stochastic Gradient Descent (SGD) algorithm for solving the matrix completion problem on a high-performance computing platform. We propose a hybrid parallel decentralized SGD framework with asynchronous inter-process communication and a novel flexible partitioning scheme to attain scalability up to hundreds of processors. We utilize Message Passing Interface (MPI) for inter-node communication and POSIX threads for intra-node parallelism. We tested our method by using different real-world benchmark datasets. Experimental results on a hybrid parallel architecture showed that, compared to the state-of-the-art, the proposed algorithm achieves 6× higher throughput on sparse datasets, while it achieves comparable throughput on relatively dense datasets.

Design and Evaluation of Deep Learning Models for Continuous Acute Pain Detection Based on Phasic Electrodermal Activity.

The current method for assessing pain in clinical practice is subjective and relies on self-reported scales. An objective and accurate method of pain assessment is needed for physicians to prescribe the proper medication dosage, which could reduce addiction to opioids. Hence, many works have used electrodermal activity (EDA) as a suitable signal for detecting pain. Previous studies have used machine learning and deep learning to detect pain responses, but none have used a sequence-to-sequence deep learning approach to continuously detect acute pain from EDA signals, as well as accurate detection of pain onset. In this study, we evaluated deep learning models including 1-dimensional convolutional neural networks (1D-CNN), long short-term memory networks (LSTM), and three hybrid CNN-LSTM architectures for continuous pain detection using phasic EDA features. We used a database consisting of 36 healthy volunteers who underwent pain stimuli induced by a thermal grill. We extracted the phasic component, phasic drivers, and time-frequency spectrum of the phasic EDA (TFS-phEDA), which was found to be the most discerning physiomarker. The best model was a parallel hybrid architecture of a temporal convolutional neural network and a stacked bi-directional and uni-directional LSTM, which obtained a F1-score of 77.8% and was able to correctly detect pain in 15-second signals. The model was evaluated using 37 independent subjects from the BioVid Heat Pain Database and outperformed other approaches in recognizing higher pain levels compared to baseline with an accuracy of 91.5%. The results show the feasibility of continuous pain detection using deep learning and EDA.

Cross Pyramid Transformer makes U-net stronger in medical image segmentation

Accurate auto-medical image segmentation, which is essential in disease diagnosis and treatment planning, has been significantly prospered by recent advances in Convolution Neural Network (CNN) and Transformer. However, pure CNN-based or pure Transformer-base architectures exhibit limitations for segmentation tasks. For example, CNN fails to capture long-range relations due to fix receptive fields, and Transformer ignores pixel-level spatial details of features. To address these challenges, we propose a novel parallel hybrid architecture for medical image segmentation, which is named CPT-Unet (Cross Pyramid Transformer U-shape Network). CPT-Unet may be the first attempt to exploit both the advantages of Pyramid Vision Transformer (PVT) and CNN to the full by integrating them into the standard U-shape network to improve the segmentation performance and inference time. Specifically, we design a parallel dual branch encoder and decoder in CPT-Unet that consists of CNN and PVT. The input image is fed into the encoder of these two branches simultaneously to extract low-level spatial details and global contexts in a much shallower way. We design a novel fusion strategy to adequately utilize the multi-scale features extracted from CNN and PVT. We also add PVT in the decoder to better restore the segmentation map. Experiments on two public segmentation datasets demonstrate the improved performance of the proposed CPT-Unet over the comparison methods. The source code is available at https://github.com/ShengYue007/CPT-Unet.git.

Development of Monte Carlo hybrid parallel algorithm and application in high resolution spatial-energy flux distribution calculation

Monte Carlo transport calculations for fusion reactors are very challenging due to factors such as large radiation decay gradients, complex geometric structures, and strong neutron anisotropy. For such deep penetration shielding calculation problems, in order to obtain accurate global flux distribution, variance reduction methods and parallel computing are necessary means. When using the global weight window variance reduction technique, the huge deviation of the particle arrival weight and the weight window can lead to an extreme number of splits, which is the long history problem. If a process encounters the long history problem in parallel computing and takes a lot of time to complete, other processes will stop computing because the histories allocated in the computing cycle have been completed, seriously degrading computing efficiency. In addition, the calculation of the shutdown dose rate, a key issue in fusion neutronics calculations, requires a high “space-energy” resolution, which brings huge challenges to computational memory, especially in parallel computing. In order to overcome the degradation of parallel efficiency caused by the long history problem and the memory limitation in large-scale shutdown dose rate calculation, a hybrid parallel algorithm based on shared memory was researched. Finally, the hybrid parallel architecture was built in the Monte Carlo code cosRMC to realize dynamic load balancing between threads. Through the memory sharing among all threads of a single node, the pressure of memory consumption can be effectively relieved. In order to verify the effectiveness of the hybrid parallel algorithm developed in this paper, it was applied to the calculation of CFETR. The results showed that the hybrid parallel algorithm can effectively alleviate the impact of the long history problem on the parallel efficiency. In terms of saving memory, the hybrid parallel algorithm also showed a significant effect, which provides a guarantee for large-scale shutdown dose rate calculation.

HybriD-GM: A Framework for Quantum Computing Simulation Targeted to Hybrid Parallel Architectures

This paper presents the HybriD-GM model conception, from modeling to consolidation. The D-GM environment is also extended, providing efficient parallel executions for quantum computing simulations, targeted to hybrid architectures considering the CPU and GPU integration. By managing projection operators over quantum structures, and exploring coalescing memory access patterns, the HybriD-GM model enables granularity control, optimizing hardware resources in distributed computations organized as tree data structures. In the HybriD-GM evaluation, simulations of Shor's and Grover's algorithms achieve significant performance improvements in comparison to the previous D-GM version, and also with other related works, for example, LIQUi|⟩ and ProjectQ simulators.

Modelling and simulation of detailed vehicle dynamics for development of innovative powertrains

Simulation with a basic representation of longitudinal vehicle dynamics is known to be sufficient for initial powertrain development activities related to efficiency and emissions such as concept application, optimal sizing, analysing the effects of physical and functional changes and also for defining basic control laws. However, when it comes to comprehensive analysis for efficiency improvement, minimizing instantaneous emission peaks or studying the impact of the new concepts on road safety, drivability and performance, the significance of detailed vehicle dynamics cannot be ignored. The work presented in this article defines a longitudinal vehicle dynamic modelling approach considering important characteristics such as the influence of normal load transfer on the varying grip of the front and rear wheels, the effect of wheel slip, and a complete representation of resistances encountered against vehicle motion with the objective of taking the analysis even closer to the actual driving conditions. The behaviour of this combined simulation platform under normal and extreme driving conditions seems to precisely follow the real scenario. This approach is a first step towards future analysis, optimization and controls development for improving transient powertrain aspects such as maximizing regenerative braking under heavy deceleration or optimizing road charging in P4 parallel hybrid architecture by managing wheel slip losses.

An Optimized Parallel Hybrid Architecture for Cryptocurrency Mining

One of the requirements for mining cryptocurrency (Crypto) is that the secured ledger of the blockchain must be updated. However, updating the secured ledger requires that the miner develop and solve complex mathematical equations in higher orders hexadecimal 64-digit solution called a hash. In addition to this challenge, the mining processes of cryptocurrency are both resource and cost-intensive. The resources required include, but not limited to mining software, hardware, power (energy usage), CPU or compute cycles, and NP-hard problem. Apart from these numerous challenges that are associated with mining cryptocurrency, the amount of speed that is required to mine a single block is core. Cryptocurrency mining speed requirement is significantly important because only miners that can have the fastest mining device are most likely to get the reward (profit) from competing for a block. In this paper, we designed and implemented a model to speed up mining process which is capable of giving miners an advantage to arrive at a block earlier. The novelty of our architecture is that our design is based on high performance computing paradigm where we achieve processor speed up by parallelizing the number of processors p. We experimented by varying p = 4, 8, 16. Our experimental results where we used the MC6800 simulated on Easy68k emulator demonstrate feasibility of our proposed model and prove that speed was an essential key to cryptocurrency mining. Keywords: Cryptocurrency, Blockchain, Architecture, Mining, Speedup, Bitcoin. CISDI Journal Reference Format Allenotor, D. & Oyemade, D. A. (2021): An Optimized Parallel Hybrid Architecture for Cryptocurrency Mining. Computing, Information Systems, Development Informatics & Allied Research Journal. Vol 12 No 1, Pp 95-104 DOI - https://doi.org/ 10.22624/AIMS/CISDI/V12N1P10. Available online at www.isteams.net/cisdijournal

Parallelization of the K-Means++ Clustering Algorithm

K-means++ is the clustering algorithm that is created to improve the process of getting initial clusters in the K-means algorithm. The k-means++ algorithm selects initial k-centroids arbitrarily dependent on a probability that is proportional to each data-point distance to the existing centroids. The most noteworthy problem of this algorithm is when running happens in sequential mode, as this reduces the speed of clustering. In this paper, we develop a new parallel k-means++ algorithm using the graphics processing units (GPU) where the Open Computing Language (OpenCL) platform is used as the programming environment to perform the data assignment phase in parallel while the Streaming SIMD Extension (SSE) technology is used to perform the initialization step to select the initial centroids in parallel on CPU. The focus is on optimizations directly targeted to this architecture to exploit the most of the available computing capabilities. Our objective is to minimize runtime while keeping the quality of the serial implementation. Our outcomes demonstrate that the implementation of targeting hybrid parallel architectures (CPU & GPU) is the most appropriate for large data. We have been able to achieve a 152 times higher throughput than that of the sequential implementation of k-means ++.

Comparative study of the energy efficiency of diesel and hybrid powered light aircraft

In recent times many research efforts are aimed at the development of aircraft hybrid propulsion technology as a means to minimize the aircraft environmental impact. The objective is to provide an additional assessment of the energy efficiency of the hybrid powered light aircraft in comparison with current state of the art Diesel powered aircraft. Typical 4-seat propeller airplane is chosen as a prototype and its flight performance model is created using open source data. The comparison is performed for a realistic flight profile including the power demands for go-around and flight to alternate destination airport. Parallel hybrid architecture is considered. A parametric study with variation of the mass, efficiency and energy capacity parameters of the hybrid power train is performed in order to determine the parameter regions where the hybrid power train has advantage over the Diesel one. The present study is based on the previous experience of the authors in assessing the energy efficiency of alternative power sources for the aviation.

An Assessment of Low-Cost Tractor Motorization with Main Farming Implements

The objective of the present work is to evaluate the performance of a low-cost tractor equipped with a parallel hybrid engine, which was simulated using AMESim software. The tractor was evaluated with three different farming implements attached to the tractor, and each implement requires a different type of power. The first simulation was executed without any implements attached. The tractor was able to run for 170 s with the electric motor only, which resulted in fuel savings during this period. The first implement, a moldboard plow, was attached for the second round of evaluation, and the electric motor ran by itself for 150 s, which also led to fuel savings during operation. During the third simulation, the tractor was attached to a Bette Harvest, which has a very high-power demand. The obtained results show that both engines were engaged to provide the required energy. During the final round of evaluation, simulations were run for a straw tub grinder. In this simulation, the electric motor ran alone until the battery was fully discharged. Thereafter, the combustion engine was activated in order to facilitate operations and to charge the battery. The results show that the parallel hybrid architecture employed for the low-cost tractor significantly decreased the CO2 emissions and minimized the consumption of fuel.

Optimization of The Through-The-Road Parallel Hydraulic Hybrid Drive Train

The through-the-road (TTR) hydraulic hybrid parallel architecture uses regenerative braking system that is placed at the rear wheels. This architecture offers a practical solution for an efficient passenger vehicle because it does not interfere with the conventional drive train. Thus, unlike other hybrid vehicles, this architecture allows the use of hybrid technology on an existing front-wheel-drive vehicle. This paper discusses the development of the architecture that involves the hydraulic circuit design and the optimization of the components size. The study simulates 972 vehicle models on all four segments on Worldwide Harmonized Light Vehicle Test Procedure (WLTP) drive cycle. The simulation finds the suitable model uses a 25cc/rev hydraulic pump, a 50 cc/rev hydraulic motor and a 10-liter hydraulic accumulator. The chosen combination allows the pump and the motor to operate within their limits, and maximizes the use of regenerative braking throughout the full WLTP drive cycle.

A hybrid MPI-CUDA approach for nonequispaced discrete Fourier transformation

Nonequispaced discrete Fourier transformation (NDFT) is widely applied in all aspects of computational science and engineering. The computational efficiency and accuracy of NDFT has always been a critical issue in hindering its comprehensive applications both in intensive and in extensive aspects of scientific computing. In our previous work Yang et al. (2018), a CUNFFT method was proposed and it shown outstanding performance in handling NDFT at intermediate scale based on CUDA (Compute Unified Device Architecture) technology. In the current work, we further improved the computational efficiency of the CUNTTF method using an efficient MPI-CUDA hybrid parallelization (HP) scheme of NFFT to achieve a cutting-edge treatment of NDFT at super extended scale. Within this HP-NFFT method, the spatial domain of NDFT is decomposed into several parts according to the accumulative feature of NDFT and the detailed number of CPU and GPU nodes. These decomposed NDFT subcells are independently calculated on different CPU nodes using a MPI process-level parallelization mode, and on different GPU nodes using a CUDA thread-level parallelization mode and CUNFFT algorithm. A massive benchmarking of the HP-NFFT method indicates that this method exhibits a dramatic improvement in computational efficiency for handling NDFT at super extended scale without loss of computational precision. Furthermore, the HP-NFFT method is validated via the calculation of Madelung constant of fluorite crystal structure, and thereafter verified that this method is robust for the calculation of electrostatic interactions between charged ions in molecular dynamics simulation systems. Program summaryProgram title: HP-NFFTCPC Library link to program files:http://dx.doi.org/10.17632/ys2y92jkwy.1Licensing provisions: GNU General Public License 3Programming language: MPI, C, and CUDA CSupplementary material: The program is designed for effective computation of large-scale nonequispaced discrete Fourier transformation (NDFT), which runs on particular computers equipped with NVIDIA GPUs. It has been tested on (a) one single computer node with Intel(R) Core(TM) i7-3770 @ 3.40 GHz (CPU) and GTX 980 Ti (GPU), and (b) MPI parallel computer nodes with the same configurations.Nature of problem: For NDFT, the computation is extremely time-consuming in many domains of computational physics due to the failure to utilize FFT directly, which often affects the computational efficiency of whole system seriously. Although the parallel method CUNFFT based on GPU has achieved a qualitative leap compared with previous methods in NDFT computation, the computation capability is limited to the throughput capacity of GPU when the size of NDFT system is big enough.Solution method: We constructed a hybrid parallel architecture, in which CPU and GPU are combined to accelerate the NDFT computation effectively. Firstly, the NDFT system is divided into several subcells via domain-decomposition method. Then MPI (Message Passing Interface) is used to implement the CPU-parallel computation with each computer node corresponding to a particular subcell, and furthermore each subcell in one computer node will be executed in GPU in parallel efficiently. In this hybrid parallel method, the most critical technical problem is how to parallelize a CUNFFT in the parallel strategy, which is conquered effectively by deep-seated research of basic principles and some algorithm skills.Restrictions: The HP-NFFT is mainly oriented to, and has shown significant computational efficiency to deal with large-scale NDFT computations. However, for a small NDFT system containing less than 10^6 particles, the mode of multiple computer nodes has no apparent efficiency advantage or even lower efficiency compared with the mode of a single computer node because of the serious network delay.

Effectiveness of hybrid powertrains to reduce the fuel consumption and NOx emissions of a Euro 6d-temp diesel engine under real-life driving conditions

Recent investigations demonstrated that the real-world driving conditions differ from those proposed in the homologation cycles. This provokes that the emissions levels in real-life conditions exceed the normative values, as shown in the recent scandal related to the NOx emissions from the passenger cars equipped with diesel engines. On the other hand, the upcoming emissions regulations will limit the CO2 emissions to very low levels, which demands a further optimization of the existing technology. One way to reduce the NOx and CO2 emissions is by electrifying the powertrain in a certain degree. The objective of this work is to evaluate the potential of implementing a parallel (P2) hybrid architecture in a compact car (class C) equipped with a diesel 1.6 Euro 6d-temp engine to reduce the emissions and fuel consumption in homologation and real-life driving cycles. This has been done using a 0D numerical vehicle model and the experimental engine maps of fuel consumption and emissions measured at steady state conditions. After that, the transient conditions were simulated in homologation cycles and real-life driving cycles measured by the authors in Spain. The numerical model was validated against experimental tests carried out in an active engine test bench, evidencing differences below 4% under the worldwide harmonized light vehicles test cycle (WLTC). In the real-life cycles, the hybridization of the powertrains improves the fuel consumption for all types of driving cycles (urban, combined and highway). The major benefits are obtained in urban driving cycles, with gains up to 50% in fuel consumption and CO2 emissions. In addition, the improvements in real-life conditions are higher than in the urban phase of the homologation cycles. On the contrary, combined real-life cycles (urban + rural + highway) show lower benefits than the homologation cycle. This is due to different energy management strategy that needs to be adapted to each driving situation. Lastly, it was found that, contrarily to the case of the homologation cycle, the NOx emissions are not reduced with the hybridization of the diesel powertrain in real-life conditions. Thus, to achieve 2021 CO2 target (95 g/km) and to reduce aftertreatment systems in diesel engines, other vehicles technologies need to be added to the full hybridization.

A Comparative Study on Coaxial Parallel and Coaxial Parallel-serial Hybrid Architectures for Plug-in Hybrid Electric Buses

This paper presents a comparative study on coaxial parallel hybrid (PH) and coaxial parallel-serial hybrid (PSH) powertrain architectures for plug-in hybrid electric buses (PHEB) in the perspective of economical performance. To do so, parameters of main components, including the engine power, the electric motor power, the capacity of battery are selected based on the same dynamic specifications of a PHEB. Then the PH electric bus (PHB) and the PSH electric bus (PSHB) are modeled in Matlab/Simulink, respectively. The dynamic programming (DP) algorithm and the nonlinear model predictive control (NMPC) based energy management controllers are designed to evaluate the fuel consumption and the energy cost of batteries. The simulation results reveal that the electric motors in PSHB possess about 10% higher equivalent efficiency and the PSHB has the superiority of energy saving effect over PHB.

Mission Analysis and Component-Level Sensitivity Study of Hybrid-Electric General-Aviation Propulsion Systems

The system-level capabilities and component-level sensitivities of hybrid-electric propulsion systems were analyzed by modeling a twin-engine general-aviation aircraft. The flight-performance model was developed using performance variables for the Tecnam P2006T found in published papers authored by the aircraft manufacturer. Both parallel and series hybrid-electric drivetrains were integrated into the aircraft performance model, and performance data were produced for various missions, degrees of electrification, battery specific energy densities, and electric motor power densities. The results quantified the improvements in battery specific energy density and electric motor power density necessary to make specific mission ranges feasible for several variants of each hybrid architecture. It was found that current technology allows a parallel hybrid configuration to achieve a maximum theoretical range of approximately 175 n mile. The results also indicated that parallel hybrid architectures will offer an effective near-term configuration, by offering greater range performance than a series hybrid with incremental future advancements in battery specific energy density and electric motor power density. However, distant future advancements in these technologies will allow series-hybrid architectures to produce similar range capabilities with improved fuel economy over parallel-hybrid architectures.

A Novel Four-Wheel-Drive Hybrid Electric Sport Utility Vehicle with Double Planetary Gears

Hybrid technology must be applied to Sport Utility Vehicles (SUVs) to meet the stringent fuel economy and emission regulations in the near future. Current hybrid SUVs realize four-wheel-drive (4WD) function which could improve its drivability and handing stability, by adding a motor to the rear axle of conventional series, parallel and power-split hybrid architectures. However, the extra motor incur more cost for the hybrid powertrain. In this paper, a novel hybrid powertrain with double outputs is proposed for 4WD SUVs, which uses a pair of planetary gears to couple the gasoline engine, two electric machines, and both front and rear wheels. The dynamics of this hybrid powertrain is built up and analyzed firstly, then Dynamic Programming (DP) is used to calculate the optimal controls and best fuel economy. Results indicates this 4WD hybrid powertrain shows greater fuel economy improvement than conventional 4WD system.