IMSL CNL - IMSL函数库

IMSL C 数值程序库 (CNL) 提供进阶的数学与统计程序库功能,让程序设计师能够在利用今日主流的程序语言C/C++开发应用程序时直接嵌入。这些支援 thread safe 的程序皆是依据知名的 IMSL Fortran 程序库的算法开发而成。

The IMSL Numerical Libraries have been the cornerstone of high-performance and desktop computing applications in science, technical and business environments for well over three decades. These embeddable mathematical and statistical algorithms are used in a broad range of applications -- including programs that help airplanes fly, predict the weather, enable innovative study of the human genome, predict stock market behavior and provide risk management and portfolio optimization. The algorithms embody the combination of High Performance Computing and High Productivity Computing.Significant benefits can be realized with the product’s ability to accelerate development time, reduce coding hassle, improve quality, and reduce development costs.

Embeddable Mathematical and Statistical Functionality

The IMSL Libraries are a comprehensive set of mathematical and statistical functions that programmers can embed into their software applications. The libraries save development time by providing pre-written mathematical and statistical algorithms that can be embedded into C, C# for .NET, Java™ and Fortranapplications, enhancing return on investment and programmer productivity. The IMSL Libraries can also be used from Python using PyIMSL Studio or the PyIMSL wrappers. Beyond choice of programming language, the IMSL Libraries are supported across a wide range of hardware and operating system environments including Windows, Linux, Apple and many UNIX platforms.

The IMSL Libraries and support services emphasize user productivity and cost-effectiveness providing asignificant return on investment by saving up to 95% of the time and cost of developing numerical algorithms.

Functional areas included in the IMSL Numerical Libraries:

Mathematics Statistics

Matrix Operations

Linear Algebra

Eigensystems

Interpolation & Approximation

Numerical Quadrature

Differential Equations

Nonlinear Equations

Optimization

Special Functions

Finance & Bond Calculations

Genetic Algorithm

Basic Statistics

Time Series & Forecasting

Nonparametric Tests

Correlation & Covariance

Data Mining

Regression

Analysis of Variance

Transforms

Goodness of Fit

Distribution Functions

Random Number Generation

Neural Networks

Benefits of Embedding the IMSL® Libraries in Your Analytic Applications

Accelerate Development

Analytical building blocks eliminating the need to write code from scratch

Numerical algorithms are developed, tested, documented, and ready to go

Save up to 95% of the time required to research and develop algorithms

Consistent commercial quality interfaces improve developer productivity

Develop Better Software Applications

You don't have to worry about coding and testing the numerical algorithms

Free up your developers' bandwidth for critical application-specific feature development

Develop Flexible Software Applications

The IMSL Libraries are written in the standard languages of C, C#, Java, and Fortran

Embed numerical analysis algorithms seamlessly into existing solutions

Applications built in C++, Python or any .NET language can easily reference the IMSL Libraries

Improve Quality and Reduce Uncertainty

The IMSL Libraries simplify your projects

A simpler project means a more predictable development and QA schedule

All IMSL algorithms are fully tested and qualified against proven testing criteria

QA efforts can focus on core application testing, not algorithm testing

The IMSL Libraries are fully documented and supported

Reduce Costs

The IMSL Libraries save up to 95% of algorithm development costs

The IMSL Libraries eliminate many hidden costs associated with algorithm development and support:

Background research

Debugging and QA

Porting to your specific environment

Documentation

Maintenance

Scaling for larger deployments

back to top

IMSL® Library Return on Investment (ROI)

Numerical Analysis Algorithm Development: More costly than you may think 

Direct Time and Cost to Develop In-House

To just research and develop one numerical analysis algorithm, it could take an effort of about 8 weeks.

Indirect Time and Cost to Develop In-House

In addition to direct costs for development, there are additional hidden costs associated with building your own routines, including:

Maintenance

Porting

Testing and Quality Assurance

Documentation

Indirect development efforts could require approximately 16 weeks of work.

This represents a total of 24 weeks to develop, test, port, and document one numerical analysis algorithm. Compare this to no more than five days to call, embed, and test an IMSL algorithm.

This comparison represents over 95% savings in time and cost!

Consider the opportunity costs as your top developers spend their time reinventing the wheel instead of using their domain knowledge to further develop your core application.

Finally, who will provide support as developers move on?

应用领域 

☆ 金融服务服务业中的投资组合方法

☆ 保险产业中之管理

☆ 存货管理与需求预测

☆ 能计算之仿真与模式建构

☆ 医学与生物科技系统研发与建构

☆ ISV 软件开发商开发应用软件中内嵌数学计算引擎

☆ IMSL C 数值程序库 6.0 提供更多函数，诸如的线性规划函数 (Linear Programming) 、神经网络预测与微分方程函数。除了这些新增功能之外，IMSL C 数值程序库在目前广泛被使用的平台上皆经过调校验证过兼容性、数值正确性与效能。

藉由功能的 IMSL C# 数值程序库打造功能的商业分析工具

商业决策：商业分析工具，而为达成且正确的决策，通过新功能设计的 IMSL C# 数学、统计程序库之使用，可创造出您所属的决策支援工具，帮助您发现、了解以及洞悉并进而能够掌握影响内部的种种变因。

运用 IMSL C# 数值程序库，您的应用程序可轻易于 Microsoft .NET 的结构中，开发属于您中的分析工具， IMSL C# 程序库功能的来源在于，其核心基础中了常用且的分析演算法，例如：求取利润的较大化（profit maximization）、产品设计（product design optimization）、供应链效能的演算（supply chain efficiency optimization）以及需求预测（demand forecasting）等。

常见处理器架构： x86-32 bit, x86-64 bit (Xeon-64, Opteron, EM64T), Itanium2, Sun UltraSPARC, IBM Power 等等

广泛的操作系统： Linux, Unix, Windows

广泛的编译器： gcc, Intel, Microsoft, Sun, IBM

商界的数值分析及数据视觉应用

JMSL 程序库的收集了数学、统计、财务、数据探勘(data mining)，以及绘图的类别 (Classes)，它是由  Pure Java 撰写。JMSL 是目前市面上一套结合数学、统计以及绘图的 Java 程序集。

身为现今的决策者，在压力下做出既且具前瞻性的决策。也因此，他们需要一套比过去更更具时效性的分析工具，能够提供他们可信任的运算输出结果，帮助他们做出自信的决策。

拥有类神经网络(Neural Network)技术的进阶预测功能

JMSL 程序库了类神经网络的技术，在现行的数据探勘、预测技术中增加了选择。这些类神经网络的预测类别，提供业界在使用历史数据建构预测模块以及对模块作时的功能。

类神经网络预测类别大的在于可以仿真人类解决问题的过程，把从历史数据中获得的知识应用到新的问题中，并且调整预测的正确性。透过这个功能，可以从历史资料中，例如将过去的成本数据应用到类神经网络中来预测未来的成本。

类神经网络的训练应用整合了数据探勘的四个步骤：数据前处理(data preprocessing)、网络训练(network training)、网络预测(network forecasting)以及数据后处理(data post processing)。

应用的统计功能在应用技术中

JMSL数值程序库了新的统计算法，协助统计学家以及分析人员来分析大量且复杂的数据，例如在生物科技以及生命科学中的应用。

JMSL 程序库中加入的部份了 heat map 图形，用在一些生命科学的应用，利用颜色来呈现二维的数组。Heat map 图形加上原来 JMSL 广泛的图形类别，帮助分析人员利用 Java 来开发进阶的数值分析应用。