sample_function: Sample functions for testing purposes

In marcianito/ppso: Particle Swarm Optimization and Dynamically Dimensioned Search, optionally using parallel computing based on Rmpi

Description

Just demo functions that can be optimized for testing purposes.

Usage

sample_function   (param_vec, maxwaittime = 0)
sample_function2  (param_vec, maxwaittime = 0)
ackley_function   (param_vec, maxwaittime = 0)
griewank_function (param_vec, maxwaittime = 0)
rastrigin_function(param_vec, maxwaittime = 0)

Arguments

param_vec	parameter vector
maxwaittime	when not set, the function is evaluated at maximum speed, otherwise a random delay of up to maxwaittime is introduced. Useful for simulating heterogenous clusters.

Value

Function value as defined by the respective function.

Note

Minimum values: sample_function: -2*dimensions sample_function2: -5 Ackley: -20-e Griewank: 0 Rastrigin: -dimensions

Author(s)

Till Francke

References

Ackley, D. H. (1987) A Connectionist Machine for Genetic Hillclimbing, Kluwer, Norwell, MA.

Griewank, A. O. (1981) Generalized descent for global optimization, Journal of Optimization Theory and Applications, 34, 11-39.

Rastrigin, L. A. (1974) Systems of Extremal Control, Nauka, Moscow. in Russian.

Examples

##sample_function is currently defined as
sample_function <-
function(param_vec,maxwaittime=0)          #objective function to be minimized
{
  obj=0
  for (i in 1:NROW(param_vec))
    obj=obj+(cos(param_vec[i])+1)
  obj=-abs(obj)

 # r <- sqrt((param_vec[1]*4)^2+(param_vec[2]*4)^2) 
#  obj=-5 * sin(r)/r 

#  if (runif(1,0,1)>0.95)
#  {
#    cat("error in objective function simulated by slave ",mpi.comm.rank(), file=paste("slave",mpi.comm.rank(),".log",sep=""))
#    error("error in objective function simulated by slave ",mpi.comm.rank())
#  }

  if (maxwaittime>0)  #function to be run on slaves, with random delay
  {
    starttime=Sys.time()
    waittime= runif(1,0,maxwaittime)
    while (as.numeric(Sys.time()-starttime)<waittime)
    {
    }
  }
  
  return(obj)
}



## sample_function2 is currently defined as
sample_function2 <-
function(param_vec,maxwaittime=0)          #objective function to be minimized
{
  r <- max(1e-320,sqrt((param_vec[1]*4)^2+(param_vec[2]*4)^2)) 
  obj=-5 * sin(r)/r 

  if (maxwaittime>0)  #function to be run on slaves, with random delay
  {
    starttime=Sys.time()
    waittime= runif(1,0,maxwaittime)
    while (as.numeric(Sys.time()-starttime)<waittime)
    {
    }
  }
  return(obj)
}

marcianito/ppso documentation built on May 30, 2019, 2:16 p.m.