PSOt - a particle swarm optimization toolbox for use with Matlab

A Particle Swarm Optimization Toolbox (PSOt) for use with the Matlab scientific programming environment has been developed. PSO is introduced briefly and then the use of the toolbox is explained with some examples. A link to downloadable code is provided. I. GENERAL INFORMATION A. Particle Swarm Optimization Toolbox (PSOt), Summary of Included Files Main files: 1) PSO – finds min/max of arbitrary MISO functions using PSO 2) trainPSO – Function to train neural nets using PSO 3) tpso1,2,3 – called by trainPSO for training 0 hidden layer, 1 hidden layer, and 2 hidden layer ANNs respectively. Support files: 4) wrapmat – convert vector(s) into matrix(ces) 5) unwrapmat – convert any 2d matrix(ces) into a row vector(s) 6) normalize – takes a 2D matrix and reformats it to any specified range 7) goplotpso – called by trainpso for graphical display during training Demo and miscellaneous files: 8) DemoTrainPSO – shows an example of using PSO to train a neural net to the XOR function 9) f6 – Schaffer’s f6 function 10) hypara – 4 dimensional hyperbolic paraboloid function 11) hypara2 – 8 dimensional hyperbolic paraboloid function The toolbox may be downloaded freely as a self extracting zipfile from: http://www4.ncsu.edu/~bkbirge/PSO/PSOt.exe