Computational advances in solving Mixed Integer Linear Programming problems To Professor Sauro Pierucci for leadership in Process Systems Engineering
Computational advances in solving Mixed Integer Linear Programming problems To Professor Sauro Pierucci for leadership in Process Systems Engineering
复制标题
解决混合整数线性规划问题的计算进展感谢 Sauro Pierucci 教授在过程系统工程领域的领导地位
DOI：
--
复制
发表时间：
2011
期刊：
影响因子：
--
通讯作者：
I. Grossmann
中科院分区：
文献类型：
--
作者：
Ricardo M. Lima;I. Grossmann
在本文中，我们确定了一些有助于有效解决混合整数线性规划（MILP）问题的计算进步。最近在算法级别和硬件级别添加到 MILP 求解器的功能有助于更有效地解决更困难和更大的问题。因此，我们将重点关注一个商业求解器中硬件和算法方面的主要进展，以展示最新功能的优点和缺点。特别关注多线程的利用、并行化模式以及启发式与分支和剪切算法的集成。用两个问题来展示一些新选项的优点。结果表明，虽然一些新功能可能有助于解决困难问题，但它们也会降低相对简单问题的性能。
In this paper, we identify some of the computational advances that have been contributing to the efficient solution of mixed-integer linear programming (MILP) problems. Recent features added to MILP solvers at the algorithmic level and at the hardware level have been contributing to the increasingly efficient solution of more difficult and larger problems. Therefore, we will focus on the main advances in terms of hardware, and algorithms in one commercial solver to demonstrate the advantages and disadvantages of the recent features. Special attention is given to the utilization of multiple threads, parallelization modes, and the integration of heuristics with the branch and cut algorithm. Two problems are used to show the advantages of some of the new options. The results show that while some of the new features may help on the solution of difficult problems, they can also reduce the performance on relatively easy problems.