An Open and Scalable Stochastic Simulation Library for Biology

描述(申请人提供)：我们的目标是填补当今生物医学模拟软件中一个令人惊讶和明显的空白：缺乏可从各种编程语言使用并能够利用各种硬件加速选项的随机模拟标准库。我们将为可伸缩的随机模拟创建一个免费、高效、可移植的库，其中包括可选择的算法，从Gillesbie引入的精确算法的快速实现到最近更快但近似的算法。该库将在Windows、MacOS、Linux和FreeBSD下支持C、C、Java、Python、Perl、MatLab和数学。在保持相同的高级前端API(应用编程接口)的情况下，我们将提供适合不同硬件场景的多个后端：典型的桌面单处理器系统、多核和多处理器系统，以及基于FPGA(现场可编程门阵列)的硬件加速板。我们还将提供SBML(系统生物学标记语言)接口层，允许轻松、直接地模拟以SBML表示的模型。所有软件以及硬件设计和配置软件都将在开源GNU LGPL许可证下发布，用于研究用途。通过引入这样一个可移植的、健壮的和灵活的库，我们希望同时减少不同团队重复重新实现同一软件所浪费的努力，并提供一个基线参考，供更多高级研究人员作为新的和改进的算法和软件开发的起点。

DESCRIPTION (provided by applicant): Our aim is to fill a surprising and glaring gap in the software available today for biomedical simulation: the lack of a standard library for stochastic simulation usable from a variety of programming languages and able to capitalize on various hardware acceleration options. We will produce a free, efficient, portable library for scalable stochastic simulation, featuring selectable algorithms ranging from a fast implementation of the exact algorithm introduced by Gillespie to recent faster but approximate algorithms. The library will feature support for C, C++, Java, Python, Perl, MATLAB and Mathematica under Windows, MacOS, Linux and FreeBSD. Keeping the same high-level front-end API (application programming interface), we will provide multiple back ends suitable for different hardware scenarios: typical desktop single-processor systems, multicore and multiprocessor systems, and FPGA-based (field programmable gate array) hardware acceleration boards. We will also provide an SBML (Systems Biology Markup Language) interface layer, allowing easy, direct simulation of models expressed in SBML. All software, as well as hardware designs and configuration software, will be released under the open-source GNU LGPL license for research use. By introducing such a portable, robust and flexible library, we hope to simultaneously reduce the effort wasted on repeated reimplementation of the same software by different groups, and provide a baseline reference that more advanced researchers can use as a jumping-off point for new and improved algorithm and software development.