Scribd
Upload
145 views3 pages

StandardPSO M

This document describes a standard particle swarm optimization (PSO) algorithm. It initializes the particle positions and velocities randomly, then iteratively updates the velocities and positions based on each particle's personal best position and the neighborhood best position. It clamps velocities and reflects positions that exceed the bounds. It evaluates the fitness at each iteration and returns the best fitness values found at intervals during search.

Uploaded by

Sutha Sangapillai
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as TXT, PDF, TXT or read online on Scribd
145 views3 pages

StandardPSO M

This document describes a standard particle swarm optimization (PSO) algorithm. It initializes the particle positions and velocities randomly, then iteratively updates the velocities and positions based on each particle's personal best position and the neighborhood best position. It clamps velocities and reflects positions that exceed the bounds. It evaluates the fitness at each iteration and returns the best fitness values found at intervals during search.

Uploaded by

Sutha Sangapillai
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as TXT, PDF, TXT or read online on Scribd
You are on page 1/ 3

function evalValues = standardPSO(fhd, functionNumber, DIM, evalsPerDIM) evalValues = zeros(1,11); nextEval = 1; evalPoints = [0.01, 0.1, 0.2, 0.3, 0.

4, 0.5, 0.6, 0.7, 0.8, 0.9, 1.0]; numEvals = 0; % PSO parameters -- D. Bratton and J. Kennedy, "Defining a standard for part icle swarm optimization," IEEE SIS, 2007, pp. 120 127. popsize = 50; w = 0.72984; c1 = 2.05 * w; c2 = 2.05 * w; % % Search space parameters range = 5; %BBOB2009 range = 100; %CEC2013 xmin = -range * ones(1,DIM); xmax = range * ones(1,DIM); vmin = -range * ones(1,DIM); vmax = range * ones(1,DIM); % Random initial positions x = 2 * range * rand(popsize,DIM) - range; % zero for initial velocity -- A. Engelbrecht, "Particle Swarm Optimization: Velocity Initialization," IEEE CEC, 2012, pp. 70-77. v = zeros(popsize,DIM); % initialize personal best positions as initial positions pbest = x; pbestCosts = feval(fhd, x', functionNumber); numEvals = numEvals + popsize; % update lbest [lbest] = update_lbest(pbestCosts, x, popsize); maxevals = DIM * evalsPerDIM; maxgenerations = floor(maxevals/popsize); for generation = 2 : maxgenerations % Update velocity v = w*v + c1*rand(popsize,DIM).*(pbest-x) + c2*rand(popsize,DIM).*(lbest -x); % Clamp veloctiy oneForViolation = v < repmat(vmin,popsize,1); v = (1-oneForViolation).*v + oneForViolation.*repmat(vmin,popsize,1); oneForViolation = v > repmat(vmax,popsize,1); v = (1-oneForViolation).*v + oneForViolation.*repmat(vmax,popsize,1); % Update position x = x + v; % Reflect-Z for particles out of bounds -- S. Helwig, J. Branke, and S. Mostaghim, "Experimental Analysis of Bound Handling Techniques in Particle Swarm Optimization," IEEE TEC: 17(2), 2013, pp. 259-271 % reflect lower bound relectionAmount = repmat(xmin,popsize,1) - x; oneForNeedReflection = relectionAmount > zeros(popsize,DIM); relectionAmount = (1-oneForNeedReflection).*zeros(popsize,DIM) + oneForN eedReflection.*relectionAmount;

% clampfirst x = (1-oneForNeedReflection).*x + oneForNeedReflection.*repmat(xmin,pops ize,1); % then reflect x = x + relectionAmount; % set velocity for reflected particles to zero v = (1-oneForNeedReflection).*v + oneForNeedReflection.*zeros(popsize,DI M); % reflect upper bound relectionAmount = repmat(xmax,popsize,1) - x; oneForNeedReflection = relectionAmount < zeros(popsize,DIM); relectionAmount = (1-oneForNeedReflection).*zeros(popsize,DIM) + oneForN eedReflection.*relectionAmount; % clampfirst x = (1-oneForNeedReflection).*x + oneForNeedReflection.*repmat(xmax,pops ize,1); % then reflect x = x + relectionAmount; % set velocity for reflected particles to zero v = (1-oneForNeedReflection).*v + oneForNeedReflection.*zeros(popsize,DI M); % Update pbest, lbest, and gbest newCosts = feval(fhd, x', functionNumber); numEvals = numEvals + popsize; for index = 1:popsize if newCosts(index) < pbestCosts(index) pbest(index,:) = x(index,:); pbestCosts(index) = newCosts(index); end end [lbest] = update_lbest(pbestCosts, x, popsize); gbestCost = min(pbestCosts); if numEvals >= evalPoints(nextEval) * maxevals evalValues(1, nextEval) = gbestCost; nextEval = nextEval + 1; end end evalValues(1, 11) = gbestCost; end % Function to update lbest function [lbest] = update_lbest(costX, x, popsize) %particle 1 is neighbours with particle n=popsize sm(1, 1)= costX(1, popsize); sm(1, 2:3)= costX(1, 1:2); [cost, index] = min(sm); if index==1 lbest(1, :) = x(popsize, :); else lbest(1, :) = x(index-1, :); end for i = 2:popsize-1

sm(1, 1:3)= costX(1, i-1:i+1); [cost, index] = min(sm); lbest(i, :) = x(i+index-2, :); end % particle n=popsize is neighbours with particle 1 sm(1, 1:2)= costX(1, popsize-1:popsize); sm(1, 3)= costX(1, 1); [cost, index] = min(sm); if index==3 lbest(popsize, :) = x(1, :); else lbest(popsize, :) = x(popsize-2+index, :); end end

You might also like