ppso - Particle Swarm Optimization and Dynamically Dimensioned Search, with parallel option
RForge.net

ppso

About ppso
Install on Windows
SVN access
Download/Files
Check results
Package R docs

SVN repository access

The SVN repository for this project can be checked out using
svn co svn://svn.rforge.net/ppso/trunk ppso

To install the latest development snapshot of this package, use

install.packages("ppso",,"http://rforge.net/",type="source")

To list branches use

svn ls svn://svn.rforge.net/ppso/branches
and for tags use
svn ls svn://svn.rforge.net/ppso/tags

Most recent SVN commits


r132 | francke | 2017-05-22 14:24:02 -0500 (Mon, 22 May 2017) | 9 lines
   M trunk/R/mpi_loop.R
   M trunk/R/prepare_mpi_cluster.R
computationally relevant: fix bug for parallel dds, multiple arguments
to objective_function
- bug when calling objective_function in prepare_mpi_cluster(): wrong
argument to do.call() statements
- added additional arguments ('...') to function definitions of
prepare_mpi_cluster() and mpi_loop(), otherwise additional arguments
to objective_function cause an error

r131 | francke | 2016-11-17 08:25:55 -0600 (Thu, 17 Nov 2016) | 1 line
   M trunk/DESCRIPTION
   M trunk/R/optim_dds.R
   M trunk/R/optim_pdds_robust.R
   M trunk/R/optim_ppso_robust.R
   M trunk/R/optim_pso.R
   M trunk/R/prepare_mpi_cluster.R
   M trunk/man/optim_ppso.Rd
- computationally irrelevant: added ... for additional arguments to objective function

r130 | francke | 2016-08-01 06:20:05 -0500 (Mon, 01 Aug 2016) | 2 lines
   M trunk/DESCRIPTION
   M trunk/R/check_execution_timeout.R
   M trunk/R/prepare_mpi_cluster.R
- computationally irrelevant: improved debug messaging

r129 | francke | 2016-05-04 05:01:18 -0500 (Wed, 04 May 2016) | 1 line
   M trunk/DESCRIPTION
   M trunk/R/check_execution_timeout.R
   M trunk/R/close_mpi.R
   M trunk/R/globvars.R
   M trunk/R/mpi_loop.R
   M trunk/R/optim_pdds_robust.R
   M trunk/R/optim_ppso_robust.R
   M trunk/man/optim_ppso.Rd
for parallel mode: improve error handling, only exit when a parameter set causes an error for a second time

r128 | francke | 2016-05-04 02:39:19 -0500 (Wed, 04 May 2016) | 1 line
   M trunk/DESCRIPTION
   M trunk/R/check_execution_timeout.R
   M web/wininstall.html
potentially relevant for p*: polling intervalls of master may have been set too long, resulting in delays or apparent hangs.

r127 | francke | 2016-05-04 01:14:30 -0500 (Wed, 04 May 2016) | 2 lines
   M trunk/R/close_mpi.R
- computationally irrelevant: replaced mpi.close.slaves with mpi.quit(), as the former caused hang under Torque/OpenMPI

r126 | francke | 2016-05-03 04:48:02 -0500 (Tue, 03 May 2016) | 2 lines
   M trunk/R/check_execution_timeout.R
- computationally irrelevant: amend previous commit

r125 | francke | 2016-05-03 04:44:44 -0500 (Tue, 03 May 2016) | 1 line
   M trunk/DESCRIPTION
   M trunk/R/check_execution_timeout.R
   M trunk/R/close_mpi.R
   M trunk/R/mpi_loop.R
   M trunk/R/optim_dds.R
   M trunk/R/optim_pdds_robust.R
   M trunk/R/optim_ppso_robust.R
   M trunk/R/prepare_mpi_cluster.R
- computationally irrelevant: extended verbose commenting of slaves and master