sample_function: Sample functions for testing purposes

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

Sample functions for testing purposes

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)
}

TillF/ppso documentation built on March 11, 2024, 1:27 a.m.