China
Africa
Explore chapters and articles related to this topic
Matrices for Engineers
Published in Abul Hasan Siddiqi, Mohamed Al-Lawati, Messaoud Boulbrachene, Modern Engineering Mathematics, 2017
Abul Hasan Siddiqi, Mohamed Al-Lawati, Messaoud Boulbrachene
Computational science, also called scientific computing or scientific computation, is a growing multidisciplinary field that uses advanced computing capabilities to explain complex problems. The scientific computing programme aims to produce highly skilled computational scientists and engineers capable of applying numerical methods and critical evaluation of their results to their field. Scientific computing is introduced clearly by Heath [5].
Oop, Mpi and Parallel Computing
Published in James J Y Hsu, Nanocomputing, 2017
Both C and Fortran are popular languages for scientific computing. Modern computer languages such as Java, Perl,PhP, andPython,etc., have many important features derived from C++. It is important for any programmer to understand the concepts of Object Oriented Programming (OOP), featuring encapsulation, function overload, operator overload, data abstraction, inheritance, modularity, polymorphism, memory allocation and pointer.
Price-Performance of Computer Technology
Published in Vojin G. Oklobdzija, Digital Design and Fabrication, 2017
The performance of computer technology depends on what you want the computer to do. Computers that are used for text processing do not need floating-point operations. Scientific computing needs heavy floating-point computation and typically high memory bandwidth.
Multi-objective trajectory optimization of the 2-redundancy planar feeding manipulator based on pseudo-attractor and radial basis function neural network
Published in Mechanics Based Design of Structures and Machines, 2023
Shenquan Huang, Shunqing Zhou, Luchuan Yu, Jiajia Wang
For the high-speed running planar feeding manipulator, the smooth trajectory is beneficial to reduce the jitter and improve the repetitive positioning accuracy. It satisfies the constraint of continuous differentiability. To generate n-segment smooth polynomial curves, n constrained equations should be constructed based on the 7-time polynomial interpolation method. Each constrained equation includes eight parameters, which are coefficients of 7-time polynomial equation. However, the eight parameters in the 7-time polynomial equation cannot be completely solved by known motion parameters such as displacement, velocity, and so on. Under this circumstance, the quadratic programming strategy in Python software is adopted. It can effectively obtain the numerical solution of the polynomial coefficients in the 7-time polynomial equation. The software named Python provides many open scientific computing libraries, such as NumPy, TensorFlow, and Cvxopt. Specifically, NumPy can compute high-dimensional arrays and matrices. TensorFlow applies to the programming of various machine learning algorithms. Cvxopt is widely used in numerical computation, operational research, and quadratic programming problems.
Development and Optimisation of a DNS Solver Using Open-source Library for High-performance Computing
Published in International Journal of Computational Fluid Dynamics, 2021
Hamid Hassan Khan, Syed Fahad Anwer, Nadeem Hasan, Sanjeev Sanghi
The large datasets used in the present DNS solver makes the process time-consuming for Input/Output (I/O) operation. The management and storage of large data in scientific computing is a critical issue. The area of computational fluid dynamics deals with a different approach to store parallel I/O data. In the sequential approach, a single process reads/ writes data and distributes data to other processes. The high memory demand on a single process limits the sequential approach for large data. Each process read/write on the individual file to get control of the memory demand on a single process. The disadvantage of this approach is to manage a large number of files at the time of post-processing. Therefore, the most efficient and parallel I/O library called hierarchical data format-5 (HDF5) is used in the present work. The HDF5 library implemented layer shown in Figure 6(a) allows multiple MPI processes to read/write in a single file. The figure represents that the HDF5 library is the interface between the program and the MPI-IO. All process uses the parallel HDF5 API. The skeleton of the collective parallel HDF5 API is shown in Figure 6(b).
Novel Python-based “ all-regressor model ” application for photovoltaic plant-specific yield estimation and systematic analysis
Published in Energy Sources, Part A: Recovery, Utilization, and Environmental Effects, 2021
Humphrey Adun, Olusola Bamisile, Mustapha Mukhtar, Mustafa Dagbasi, Doga Kavaz, Ariyo Oluwasanmi
This section explains the procedure for the development of the Python-based all-regressor machine learning algorithm used in this study, as shown in Figure 2. The high-level interactivity and maturing scientific library package establish Python programming language as an appealing choice for scientific computing and the development of algorithms for data analysis (Varoquaux et al. 2015). The “scikit-learn is a broad package in python for building state-of-the-art machine learning algorithms. The aim of the “all regressor algorithm” is to develop an auto-machine learning model for any dataset (PV output prediction) that would rank the models in terms of prediction accuracy and utilize the most efficient regression algorithm, as shown in Figure 3. Optimum parameters for each regressor are set in the code development using trial and error methods. The code development process involves, importing and reading the dataset, and understanding the target variable. The distribution of the output variable is analyzed. First, a code is written to transform the input data to reduce the skewness and kurtosis. Then, by using the “dataset.isna.sum/len (dataset)”,” the columns with a null ratio higher than 60% will be dropped. The program generates a heat map for the correlation between the features and afterward. The data are split into training and validation data sets using the “train-test-split” line of code. Then, using “Sktlearn.utils.testing import all-estimators”, the 20 machine learning regression models are imported to model and estimate the target variable (which is the PV specific yield). The section of the code which generates the regressor algorithm is given below.