cplex¶
cplex是IBM公司开发的一款商业版的优化引擎,该引擎专门用于求解大规模的线性规划（LP）、二次规划（QP）、带约束的二次规划（QCQP）、二阶锥规划（SOCP）等四类基本问题，以及相应的混合整数规划（MIP）问题。
cplex安装以及使用说明¶
如下图所示，登录 IBM cplex 官网 并注册账号，然后下载linux平台下的cplex安装包(试用免费版)：
申请计算资源：
srun -p 64c512g -n 4 --pty /bin/bash (思源一号)
或者
srun -p cpu -N 1 --ntasks-per-node 40 --exclusive --pty /bin/bash (pi2.0)
加载所需模块：
module load gcc openmpi jdk python
在自己的家目录下新建一个目录cplex作为安装目录，进入该目录，并将上一步得到的安装包上传到当前目录：
mkdir cplex
cd cplex
执行以下命令使安装文件具有执行权限，然后执行该文件：
chmod 777 cos_installer_preview-22.1.1.0.R0-M08SWML-linux-x86-64.bin
./cos_installer_preview-22.1.1.0.R0-M08SWML-linux-x86-64.bin
根据终端输出的安装提示一步一步往下执行即可。
小心
一定要将默认的安装路径 /opt/ibm/ILOG/CPLEX_Studio_Community2211 改为自己家目录下的某个目录，并且使用绝对路径(比如说/dssg/home/acct-hpc/hpcpzz/cplex)；
cplex自带了很多标准算例，用户可以通过执行以下命令来编译并运行这些算例：
export PATH=$PATH:~/cplex/cplex/bin/x86-64_linux:~/cplex/cpoptimizer/bin/x86-64_linux
cd ~/cplex/cplex/examples/x86-64_linux/static_pic
make execute -j
此时在终端中可得到如下结果：
make all_c
make[1]: Entering directory '/dssg/home/acct-hpc/hpcpzz/cplex214/cplex/examples/x86-64_linux/static_pic'
gcc -c -m64 -fPIC -I../../../include ../../../examples/src/c/lpex1.c -o lpex1.o
gcc -c -m64 -fPIC -I../../../include ../../../examples/src/c/lpex2.c -o lpex2.o
gcc -c -m64 -fPIC -I../../../include ../../../examples/src/c/lpex3.c -o lpex3.o
gcc -c -m64 -fPIC -I../../../include ../../../examples/src/c/lpex4.c -o lpex4.o
gcc -c -m64 -fPIC -I../../../include ../../../examples/src/c/lpex5.c -o lpex5.o
gcc -c -m64 -fPIC -I../../../include ../../../examples/src/c/lpex6.c -o lpex6.o
gcc -c -m64 -fPIC -I../../../include ../../../examples/src/c/lpex7.c -o lpex7.o
gcc -c -m64 -fPIC -I../../../include ../../../examples/src/c/lpex8.c -o lpex8.o
gcc -c -m64 -fPIC -I../../../include ../../../examples/src/c/mipex1.c -o mipex1.o
gcc -c -m64 -fPIC -I../../../include ../../../examples/src/c/mipex2.c -o mipex2.o
gcc -c -m64 -fPIC -I../../../include ../../../examples/src/c/mipex3.c -o mipex3.o
gcc -c -m64 -fPIC -I../../../include ../../../examples/src/c/mipex4.c -o mipex4.o
gcc -c -m64 -fPIC -I../../../include ../../../examples/src/c/miqpex1.c -o miqpex1.o
...
./admipex7 ../../../examples/data/mexample.mps
Selected objective sense: MINIMIZE
Selected objective name: obj
Selected RHS name: rhs
Selected bound name: bnd
Version identifier: 22.1.1.0 | 2022-11-26 | 9160aff4d
CPXPARAM_MIP_Strategy_Search 1
Legacy callback S
Found incumbent of value -46.000000 after 0.00 sec. (0.00 ticks)
Tried aggregator 2 times.
Aggregator did 1 substitutions.
Reduced MIP has 2 rows, 3 columns, and 6 nonzeros.
Reduced MIP has 0 binaries, 1 generals, 0 SOSs, and 0 indicators.
Presolve time = 0.00 sec. (0.01 ticks)
Tried aggregator 1 time.
Reduced MIP has 2 rows, 3 columns, and 6 nonzeros.
Reduced MIP has 0 binaries, 1 generals, 0 SOSs, and 0 indicators.
Presolve time = 0.00 sec. (0.00 ticks)
MIP emphasis: balance optimality and feasibility.
MIP search method: traditional branch-and-cut.
Parallel mode: none, using 1 thread.
Root relaxation solution time = 0.00 sec. (0.02 ticks)
Nodes Cuts/
Node Left Objective IInf Best Integer Best Bound ItCnt Gap Variable B NodeID Parent Depth
* 0+ 0 -46.0000 -163.0000 254.35%
* 0+ 0 -111.0000 -163.0000 46.85%
* 0+ 0 -122.5000 -163.0000 33.06%
0 0 -125.2083 1 -122.5000 -125.2083 11 2.21%
0 0 cutoff -122.5000 11 0.00% 0 0
Elapsed time = 0.01 sec. (0.09 ticks, tree = 0.01 MB, solutions = 3)
Root node processing (before b&c):
Real time = 0.01 sec. (0.09 ticks)
Sequential b&c:
Real time = 0.00 sec. (0.00 ticks)
------------
Total (root+branch&cut) = 0.01 sec. (0.09 ticks)
Solution status 101.
Objective value -122.5
...
参考资料¶
最后更新:
2024 年 10 月 14 日