SOPT - 数学建模软件

SOPT是一款线性、非线性和整数软件。它为工业决策系统提供了多种新的算法和应用能力。基于数学编程中的突破性技术，SOPT快速解决问题。

强调：

目前SOPT有三种版本： SOPT-CP是线性和非线性的程序； SOPT-IP是线性和整数的程序； SOPT-SPSOPT-IP和AMPL。模块都具有算法的实际实现，这可以帮助您的解决方案开发解决复杂的决策问题。

对于线性程序：

• 快速，性能稳定

• 对变量或约束的数量没有限制

• 算法（例如，Newton Barrier，Simplex，Cross-over）

对于整数程序：

• 实用的启发式搜索

• 易于使用的定制参数

• 启动以进行方案分析

对于非线性程序：

• 二次

• 用于凸函数的的用户子程序

• 全局功能

平台：

Win32、Unix

SOPT可以单独使用，也可以作为子程序库或DLL使用。

新版本的SOPT

SOPT新版本为4.2，随着发布，SOPT创建了动态日志，其中有解决方案值和缩短时间的信息，以便及时SOPT仍然在解决问题，用户也可以访问正在进行的状态。此功能对于SaaS应用程序等在线应用程序有用。

新的版本从SOPT数据结构中检索多个证书解决方案以及CSV各式的解决方案文件等功能。选择多个近似解决方案中的是一种实用且方便的方法，可以为您的应用确定较好解决方案，而较好客观价值可能并不是较好的解决方案。

之前我们在版本3中引用了一种新的搜索算法。这种宽展搜索伏安法实用切割平面和子LP解决方案，它视图通过原理局部来找到更好的整数解。它已经应用于困难的组合问题，例如集合覆盖和TSP问题，这些问题经常出现在调度和路由应用中。版本还GUI改进，例如显示LP和IP解决方案性能。

【英文介绍】

SOPT is an efficient optimizer for linear, integer, and nonlinear programs. It offers a variety of new and fast algorithms and application capabilities for industrial decision support systems. Based on breakthrough technologies in math programming, SOPT solves problems with lightning speed and robustness.

Highlights

Currently SOPT has three versions, SOPT-CP for linear and nonlienar programs, SOPT-IP for linear and integer programs, and SOPT-SP (SOPT-IP and AMPL). Each module has practical implementation of unique algorithms, which may help your solution development to solve complex decision problems.

Supported Plaforms: Win32   Unix
SOPT can be used stand-alone, or as a subroutine library or DLL.

Latest version of SOPT

A new version SOPT 4.2 will be released on August 25. With theis release, SOPT creates a dynamic optimization log with information on incumbent solution values and elasped times so that users can access the ongoing optmization status even when SOPT is still solving the problem. This feature is particularly useful for online applications such as SaaS appliacations.

Recent releases include such features as retrieval of multiple integer solutions from the SOPT data structure, and various solutions files in the CSV format. Choosing one of multiple near-optiomal solutions is a practical and convenicent way to identify the best solution for your application, when the best objective value may not mean the best solution. Recent release note for SOPT is available from here.

Previously we introduced a new search algorithm in version 3. This extended search algorithm uses cutting planes and sub LP solutions, and it tries to find better integer solutions by moving away from local optima. It has been applied to difficult combinatorial problems, such as set covering and TSP problems, which are often found in scheduling and routing applications. The last version also included GUI improvement such as displaying LP and IP solution performance.