Documentation of seltrunc

Global Index (all files) (short | long) | Local contents | Local Index (files in subdir) (short | long)

Function Synopsis

NewChromIx = seltrunc(FitnV, Nsel, Dummy);

Help text

 SELection by TRUNCation

 This function performs SELection by TRUNCation.

 Syntax:  NewChromIx = seltrunc(FitnV, Nsel)

 Input parameters:
    FitnV     - Column vector containing the fitness values of the
                individuals in the population.
    Nsel      - Number of individuals to be selected

 Output parameters:
    NewChromIx- Column vector containing the indexes of the selected
                individuals relative to the original population, shuffeld.
                The new population, ready for mating, can be obtained
                by calculating OldChrom(NewChromIx,:).

 For the truncation threshold the inverse of the selection
 pressure is used (Trunc = 1/SP).
 1/SP is computed from FitnV by mean(FitnV)/max(FitnV)

 See also: selection, selsus, selrws, seltour, sellocal

 Example:
    % define fitness vector
       FitnV = [.1; .9; 1.6; 2.0; 0.4; 1.3; 1.7; 0.7; 0.2];
    % selects 6 indices from FitnV, Trunc = 0.5;
       NewChromIx = seltrunc(FitnV, 6);
    % possible result
       NewChromIx = [3; 4; 2; 4; 6; 7];
    % Get selected individuals from population Chrom
       SelChrom = Chrom(NewChromIx, :)

Cross-Reference Information

Listing of function seltrunc

% Author:     Hartmut Pohlheim
% History:    17.05.94     file created
%             11.10.95     examples added


function NewChromIx = seltrunc(FitnV, Nsel, Dummy);

% Uniform at random or fitness based selection from truncated parents
   SELPAR = 0;          % 0: uniform at random, 1: fitness based

% Identify the population size (Nind)
   [Nind, ans] = size(FitnV);

% Trunc - truncation threshold in the range [0, 1],
% computed from selection pressure 1/SP
   Trunc = (sum(FitnV) / Nind) / max(FitnV);
   if (Trunc < 0 | Trunc > 1),
      disp('Trunc in truncation selection outside range, set to 0.5');
      Trunc = 0.5;
   end

% Truncate Trunc% best individuals
   [Dummy, Ix1] = sort(-FitnV);
   Ix1 = Ix1(1:floor(Nind * Trunc));

% Assign fitness values
   if SELPAR == 0,
   % Selects parents from truncated individuals uniform at random
      FitnV1 = ones(length(Ix1), 1);
   else      % if SELPAR == 1,
   % Select parents from truncated individuals fitness based
      FitnV1 = FitnV(Ix1);
   end

% Select Nsel individuals from truncated individuals
   Ix2 = selsus(FitnV1, Nsel);

% Create index as row vector
   NewChromIx = Ix1(Ix2);
   NewChromIx = NewChromIx(:); 


% End of function