Copyright (C) 2009 A. Ismael F. Vaz and L. N. Vicente.
http://www.norg.uminho.pt/aivaz http://www.mat.uc.pt/~lnv
http://www.norg.uminho.pt/aivaz/pswarm
This library is free software; you can redistribute it and/or modify
it under the terms of the GNU Lesser General Public License as
published by the Free Software Foundation; either version 2.1 of the
License, or (at your option) any later version.
This library is distributed in the hope that it will be useful, but
WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
Lesser General Public License for more details.
You should have received a copy of the GNU Lesser General Public
License along with this library; if not, write to the Free Software
Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307,
USA.
===========================
Parallel PSwarm Version 1.5
===========================
---------------
Read me file
---------------
Welcome to Parallel Pattern Swarm version 1.5 (PPSwarm v1.5)
This code provides an implementation of the algorithm described in:
A.I.F. Vaz and L.N.Vicente, A particle swarm pattern search method
for bound constrained global optimization, Journal of Global
Optimization, 39 (2007) 197-219.
The linear constrained approach is developed in the paper:
A.I.F. Vaz and L.N. Vicente. PSwarm: A hybrid solver for linearly
constrained global derivative-free optimization.
Available online at:
www.norg.uminho.pt/aivaz/pswarm
This file describes the following topics:
1. Change log
2. System requirements
3. Compilation/Installation
4. Contents of the ppswarm_v1_1 directory
5. The command line
6. Credits
1. Change log
----------------------------------------
PSwarm v0.1 implements the algorithm described in
A. I. F. Vaz and L. N. Vicente, A particle swarm pattern search
method for bound constrained global optimization, Journal of Global
Optimization, 39 (2007) 197-219. (available online at:
http://dx.doi.org/10.1007/s10898-007-9133-5)
PSwarm v0.1 provides a standalone version (where the user must
provide a simple bound optimization problem) that can run in serial
or parallel mode.
PSwarm v0.1 provides also an AMPL (www.ampl.com) interface version
in serial mode.
Version v1.1 implements the algorithm described in
A.I.F. Vaz and L.N. Vicente. PSwarm: A hybrid solver for linearly
constrained global derivative-free optimization.
PSwarm v1.1 is now able to handles both bound and linear
constraints.
PSwarm v1.1 provides a R (www.r-project.org) and Python
(www.python.org) interface in addition to the AMPL interface.
PSwarm v1.2 corrects a bug in the pswarm_py.c file and in the
makefile.
PSwarm v1.3 allows user to provide a iteration print function called
each iter mod iprint = 0, where iter is the iteration number and
iprint is a new option available in the algorithm. These options
(with many others) are now exported to R and Python.
PSwarm v1.4 calls the objective function in a vectorized way.
PSwarm v1.5 corrects some memory allocation, allowing bigger
problems to be addressed.
2. System requirements
----------------------------------------
PSwarm v1.4 can be compiled with or without linear constraints
support.
This option is available at compilation time allowing users to still
use the PSwarm algorithm in different systems configurations.
PSwarm with only bound constraints support:
- C compiler (gcc for the serial Linux version, mpicc for the
parallel Linux version or VC++ for the serial Windows version);
- AMPL interface library for the AMPL version.
PSwarm with bound and linear constraints:
- C compiler (gcc for the serial Linux version, mpicc for the
parallel Linux version or VC++ for the serial Windows version);
- BLAS (Basic Linear Algebra Subroutines) library (may be already
available on your system, otherwise download it from the web and
make a local BLAS library);
- LAPACK (Linear Algebra PaCKage) library (may be already available
on your system, otherwise download it from the web and make a
local LAPACK library);
- AMPL interface library for the AMPL version;
- A system Python installation for the Python dll version;
- A system R installation for the R dll version.
3. Installation
----------------------------------------
The code requires approximately 12.5Mb of hard disk space
(source files, windows libraries, AMPL student edition binaries,
etc).
Unzip the PPSwarm_v1_1.zip file in any location using your favorite
unzip software. A directory PPSwarm_v1_1 will be created, having as
subdirectories 'nl', 'include', 'libs', which refers to the AMPL
problems models (in the .nl format), the corresponding interface
include files and a directory to place additional libs (AMPL, BLAS
or LAPACK), respectively.
Two makefiles are provided to assist in the solver compilation.
'makefile.vc' is for the Microsoft Windows Visual C++ and 'makefile'
is for Linux.
LINUX:
------
A makefile for Linux is provided.
In order to obtain the PSwarm with the linear constraints support a
'_linear' suffix should be added to the make command line (e.g.
'make ampl_linear' will make the pswarm binary with AMPL and linear
constraints support).
Type 'make serial' for the standalone serial version. In the
standalone version the user must provide the problem definition in
the file user.c. The 'objfun' C function is expected to return the
objective function value at 'x'. The 'set_problem_dimension' return
the number of problem variables and 'set_problem' sets the lower and
upper bounds on the variables together with an initial guess. The
user may use the 'user_init' function to initialize some data.
Type 'make parallel' for a standalone parallel version (see previous
comments on how to tune the objective function and bounds). There
is no manual for the parallel version, but the code just assigns to
each processor the computation of an objective function value. We
advice to have 2*(n+1)+1 processor (since currently we have
2*(n+1) directions in the poll step plus one processor for
running the algorithm). See file 'job.pbs' for an example on using
pswarm in a PBS system.
Type 'make ampl' to make the ampl version of pswarm.
Type 'make lib' to make the a pswarm library (libpswarm.a). The user
functions defined in user.c are not included and must be provided by
the user.
Type 'make r' to make the pswarm dynamic library to be used with R
(see the provided .r example).
Type 'make py' to make the pswarm dynamic library to be used with
Python (see the provided .py example).
WINDOWS:
--------
A makefile.vc for Microsoft VC++ is provided. The same comment
applies to the _linear suffix in the make command line arguments.
Type 'nmake -f makefile.vc serial' for the serial standalone
version.
Type 'nmake -f makefile.vc ampl' for the AMPL version.
Type 'nmake -f makefile.vc rr' for the R dynamic library. Note that
under Windows its 'rr' for the version without linear constraints
and 'r_linear' for the version with linear constraints.
Type 'nmake -f makefile.vc py' for the Python dynamic library.
Type 'nmake -f makefile.vc lib' for the pswarm library.
WINDOWS MPI version is not provided.
4. Contents of the PPSwarm_v1_5 directory
----------------------------------------
In the directory PPSwarm_v1_5 there are the following files:
------ README -------------
README.txt - This readme file
lgpl.tx
没有合适的资源?快使用搜索试试~ 我知道了~
资源推荐
资源详情
资源评论
收起资源包目录
OPTI工具包及对应的文件.rar (1152个子文件)
AUTHORS 4KB
AUTHORS 2KB
AUTHORS 2KB
AUTHORS 2KB
AUTHORS 1002B
AUTHORS 835B
AUTHORS 366B
AUTHORS 187B
AUTHORS 153B
AUTHORS 153B
AUTHORS 67B
AUTHORS 48B
AUTHORS 16B
msvc141opts.bat 3KB
msvc140opts.bat 3KB
msvc140engmatopts.bat 3KB
IEEE014.branch 2KB
IEEE014.bus 2KB
lpsolve.c 161KB
reader_nl.c 41KB
filtersdmex.c 36KB
csdpmex.c 35KB
nloptmex.c 33KB
dsdpmex.c 31KB
amplmex.c 29KB
mumpsmex.c 23KB
levmarmex.c 20KB
nl2solmex.c 20KB
mkltrnls.c 20KB
lmdermex.c 19KB
hybrjmex.c 18KB
pswarmmex.c 17KB
m1qn3mex.c 14KB
matlab.c 10KB
symb_ctemp.c 4KB
ChangeLog 8KB
changes 1KB
config 329B
COPYING 35KB
COPYING 35KB
COPYING 35KB
COPYING 18KB
COPYING 18KB
COPYING 12KB
COPYING 6KB
COPYING 2KB
COPYRIGHT 2KB
scipmex.cpp 55KB
nomadmex.cpp 50KB
cbcmex.cpp 39KB
scipnlmex.cpp 39KB
glpkcc.cpp 32KB
ooqpmex.cpp 32KB
opti_mex_utils.cpp 29KB
clpmex.cpp 23KB
scipsdpmex.cpp 22KB
callbackfunctions.cpp 21KB
callbackfunctions.cpp 19KB
options.cpp 18KB
gslmex_nls.cpp 17KB
symb_cadtemp.cpp 16KB
coinR.cpp 14KB
rmathlib.cpp 12KB
bonminmex.cpp 11KB
coinW.cpp 11KB
matlabprogram.cpp 10KB
sparsematrix.cpp 9KB
sparsematrix.cpp 9KB
ipopt.cpp 9KB
matlabprogram.cpp 9KB
lbfgsb.cpp 8KB
options.cpp 7KB
bonminoptions.cpp 7KB
lbfgsb_program.cpp 7KB
program.cpp 5KB
ipoptoptions.cpp 5KB
mklJac.cpp 5KB
matlabinfo.cpp 4KB
gslmex.cpp 3KB
iterate.cpp 3KB
iterate.cpp 3KB
matlabjournal.cpp 2KB
matlabinfo.cpp 2KB
matlabjournal.cpp 2KB
scipeventmex.cpp 2KB
matlabfunctionhandle.cpp 2KB
matlabfunctionhandle.cpp 2KB
spIndex.cpp 2KB
matlabexception.cpp 1KB
matlabexception.cpp 986B
ClpCrashSource.cpp 591B
lbti-00-40.dat 24KB
lbti-00-20.dat 16KB
lbti-00-15.dat 14KB
train1.dat 4KB
c-reload-14e.dat 3KB
c-reload-14d.dat 3KB
oil.dat 3KB
c-sched2.dat 3KB
prod03.dat 3KB
共 1152 条
- 1
- 2
- 3
- 4
- 5
- 6
- 12
资源评论
- qq_434867552019-12-11209836505@qq.com,谢谢施舍
- qq_275612952019-08-13scip 怎么没有啊?
oldsoliderneverdie
- 粉丝: 119
- 资源: 3
上传资源 快速赚钱
- 我的内容管理 展开
- 我的资源 快来上传第一个资源
- 我的收益 登录查看自己的收益
- 我的积分 登录查看自己的积分
- 我的C币 登录后查看C币余额
- 我的收藏
- 我的下载
- 下载帮助
安全验证
文档复制为VIP权益,开通VIP直接复制
信息提交成功