gRaviopt: Agent-based stochastic global optimization.

In gRaviopt: What the package does (short line)

Description

Performs stochastic global optimization via Newtons laws of gravity and motion.

Usage

gRaviopt(fn, Par, lower.limits = -10, upper.limits = 10, n = 20, m = 20, iterations = 200, man.scaling = FALSE, alpha = 0.1)

Arguments

fn	the function to be optimized (maximised). NA and NaN values are not allowed.
Par	number of parameters used in "fn".
lower.limits	the lower limits of the parameters used in "fn".
upper.limits	the upper limits of the parameters used in "fn".
n	number of agents that should be used for the optimization process.
m	number of best solutions to be stored in the memory.
iterations	number of iterations used for optimization process.
man.scaling	the user can specicify wether to choose the the radius of local optimization manually or automatically.
alpha	radius that divises between local and global optimization.

The output of the function gRaviopt contains a matrix GMemory with the m best solutions and an array GP of all calculations.

Author(s)

Peter Kehler [e-mail]

References

A. Kaveh and S. Talatahari: A novel heuristic optimization method: charged system search, Acta Mech 213, 267–289 (2010)

See Also

gRaviopt.Plot for graphical output.

Examples

## Rastrigin02 function
## The function has a global maximum f(x) = 0 at the point (0,0).  
## gRaviopt searches for maxima of the objective function between
## lower and upper bounds on each parameter to be optimized. 

Rastrigin02  <- function(X){
  -((X[,1]^2 - 10*cos(2*pi*X[,1]^2) + 10) + (X[,2]^2 - 10*cos(2*pi*X[,2]^2) + 10))
}

## This version of the function is needed for gRaviopt.Plot
Rastrigin02.2d  <- function(x,y){
  -((x*x - 10*cos(2*pi*x) + 10) + (y*y - 10*cos(2*pi*y) + 10))
}

# optimization process of Rastrigin02
Rast02 <- gRaviopt(fn= Rastrigin02, Par=2, n=20, lower.limits = -3, upper.limits = 3,man.scaling=TRUE,alpha=0.05)

# the best solutions found
Rast02$Memory

# the movements of the particles during the optimization process
gRaviopt.Plot(fn= Rastrigin02.2d, gRaviopt.Result=Rast02, Par=2, iterations=200, n=20, lower.limits = -3, upper.limits = 3, Movements=TRUE,man.scaling=TRUE,alpha=0.1,Nice=FALSE)

gRaviopt documentation built on May 2, 2019, 6:53 p.m.