[pOpt,fVal,stat] = NDBASE.PSO([],func,p0,’Option1’,’Value1’,…)
[pOpt,fVal,stat] = NDBASE.PSO(dat,func,p0,’Option1’,’Value1’,…)
PSO finds a minimum of a function of several variables using the particle swarm optimization (PSO) algorithm originally introduced in 1995 by Kennedy and Eberhart. This algorithm was extended by Shi and Eberhart in 1998 through the introduction of inertia factors to dampen the velocities of the particles. In 2002, Clerc and Kennedy introduced a constriction factor in PSO, which was later on shown to be superior to the inertia factors. Therefore, the algorithm using a constriction factor was implemented here.
The function has two different modes, depending on the first input argument. If dat is empty, pso minimizes the cost function func, that has the following header: R2 = func(p)
If dat is a struct, the function optimizes the model defined by func via least squares to the data stored in dat. In this case func has the following header: yModel = func(x,p) And the least square deviation is defined by: R2 = sum((yData-yModel).^2/errorData.^2)
Input:
dat Either empty or contains data to be fitted stored in a structure with fields: dat.x vector of N independent variables, dat.y vector of N data values to be fitted, dat.e vector of N standard deviation (positive numbers) used to weight the fit. If zero or missing 1./dat.y^2 will be assigned to each point. func Function handle with one of the following definition: R2 = func(p) if dat is empty, y = func(x,p) if dat is a struct. Here x is a vector of N independent variables, p are the M parameters to be optimized and y is the simulated model, R2 is the value to minimize. p0 Vector of M initial parameters.
Options:
lb Vector with N elements, lower boundary of the parameters. Default value is -1e5. ub Vector with N elements, upper boundary of the parameters. Default value is 1e5. MaxIter Maximum number of iterations, default value is 100M. MaxFunEvals Maximum number of function evaluations, default value is 1000M. NOT IMPLEMENTED! TolX Convergence tolerance for parameters, default value is 1e-3. Display Level of information to print onto the Command Line during optimisation. Possible values are ‘off’, ‘notify’ and ‘iter’. Default is ‘off’. TolFun Convergence tolerance on the R2 value (return value of func or the weighted least square deviation from data). Default value is 1e-3. SwarmC1 Swarm parameter, also called phi_p in the literature. Default value is 2.8. SwarmC2 Swarm parameter, also called phi_g in the literature. Default value is 1.3. k0 Swarm paramter, it can take values between 0 and 1, but is usually set to one (Montes de Oca et al., 2006). Default value is 1. seed Seed number for the random number generator, by default it is seeded from the system clock every time the function is called. PopulationSize Size of the swarm, default value is 25. autoTune Determine the size of the swarm based on the number of free parameters (number of parameters that are not fixed by setting lb(i)==ub(i)). Default is false. The optimal swarm size is: PopulationSize = 25 + 1.4*Nf This value is based on the technical report: Good Parameters for Particle Swarm Optimization By Magnus Erik Hvass Pedersen Hvass Laboratories Technical Report no. HL1001 http://www.hvass-labs.org/people/magnus/publications/pedersen10good-pso.pdf tid Setup time estimation with the following values: 0 Do nothing (default). 1 Text output to the Command Window. Default. 2 Graphical output, using the waitbar() function. -1 Set the value from the SpinW preferences (needs SpinW installed).
See also NDBASE.LM.