ECF 1.5
Classes
BBOB function minimization (from http://coco.gforge.inria.fr)

Function minimization example - using FloatingPoint or Binary genotype (see more on this example in help/tutorial.html) More...

Classes

class  FunctionMinEvalOp
 Function minimization evaluation class. More...
 

Detailed Description

Function minimization example - using FloatingPoint or Binary genotype (see more on this example in help/tutorial.html)

Description
Tries to find the (global) minimum of a continuous multidimensional function. The example uses either FloatingPoint or Binary genotype. Functions are implemented in Black-Box Optimization Benchmarking package available at http://coco.gforge.inria.fr. The location of the optimum is drawn uniformly from [-4, 4] in all dimensions for all functions. The value of the optimum of all functions is set to zero in this example (in original BBOB package the value is in [-100, 100]).

The example produces two executables (separate VS projects):

To run the ecf_experiment executable:

Parameters (in the config file)
  • "coco.function": defines the minimization function to be used (default: 1); 1-24 noiseless functions, 101-130 noisy
  • "coco.enableoutput": should COCO framework output results for post-processing (disabled by default)
  • "coco.folder": destination folder for COCO output; must be created beforehand (see /bbob/README.txt)
  • with FloatingPoint genotype, a number of dimensions and the interval [lbound, ubound] must be set
  • with Binary genotype, a number of dimensions, the interval [lbound, ubound] and precision (number of digits after the decimal point) must be set
References