jTreeSeedAlgorithm.m

%[2015]-"TSA: Tree-seed algorithm for continuous optimization"

% (9/12/2020)

function TSA = jTreeSeedAlgorithm(feat,label,opts)
% Parameters
lb    = 0;
ub    = 1; 
thres = 0.5; 
ST    = 0.1;    % switch probability

if isfield(opts,'T'), max_Iter = opts.T; end
if isfield(opts,'N'), N = opts.N; end
if isfield(opts,'ST'), ST = opts.ST; end 
if isfield(opts,'thres'), thres = opts.thres; end

% Objective function
fun = @jFitnessFunction;
% Number of dimensions
dim = size(feat,2); 
% Initial (5)
X   = zeros(N,dim); 
for i = 1:N
  for d = 1:dim
    X(i,d) = lb + (ub - lb) * rand();
  end 
end
% Fitness
fit = zeros(1,N); 
for i = 1:N 
  fit(i) = fun(feat,label,(X(i,:) > thres),opts);
end
% Best solution (6)
[fitG, idx] = min(fit);
Xgb         = X(idx,:);
% Maximum & minimum number of seed
Smax = round(0.25 * N); 
Smin = round(0.1 * N);
% Pre
curve = zeros(1,max_Iter);
curve(1) = fitG;
t = 2;
% Iteration
while t <= max_Iter
  for i = 1:N
    % Random number of seed
    num_seed = round(Smin + rand()* (Smax - Smin)); 
    Xnew     = zeros(num_seed, dim);
    for j = 1:num_seed
      % Random select a tree, but not i
      RN = randperm(N); 
      RN(RN == i) = []; 
      r  = RN(1);
      for d = 1:dim
        % Alpha in [-1,1]
        alpha = -1 + 2 * rand();
        if rand() < ST  
          % Generate seed (3)
          Xnew(j,d) = X(i,d) + alpha * (Xgb(d) - X(r,d));
        else
          % Generate seed (4)
          Xnew(j,d) = X(i,d) + alpha * (X(i,d) - X(r,d));
        end
      end
      % Boundary
      XB = Xnew(j,:); XB(XB > ub) = ub; XB(XB < lb) = lb;
      Xnew(j,:) = XB;
    end
    % Fitness
    for j = 1:num_seed
      % Fitness
      Fnew = fun(feat,label,(Xnew(j,:) > thres),opts);
      % Greedy selection
      if Fnew < fit(i)
        fit(i) = Fnew;
        X(i,:) = Xnew(j,:);
      end
    end
  end
  % Best solution (6)
  [fitG_new, idx] = min(fit);
  Xgb_new         = X(idx,:);
  % Best update
  if fitG_new < fitG
    fitG = fitG_new;
    Xgb  = Xgb_new;
  end
  % Store 
  curve(t) = fitG;
  fprintf('\nIteration %d Best (TSA)= %f',t,curve(t))
  t = t + 1;
end
% Select features
Pos   = 1:dim; 
Sf    = Pos((Xgb > thres) == 1); 
sFeat = feat(:,Sf);
% Store results
TSA.sf = Sf;
TSA.ff = sFeat; 
TSA.nf = length(Sf); 
TSA.c  = curve; 
TSA.f  = feat;
TSA.l  = label;
end