NAG C Library Function Document

nag_glopt_bnd_pso (e05sac)

Note: this function uses optional parameters to define choices in the problem specification and in the details of the algorithm. If you wish to use default settings for all of the optional parameters, you need only read Sections 1 to 10 of this document. If, however, you wish to reset some or all of the settings please refer to Section 11 for a detailed description of the algorithm and to Section 12 for a detailed description of the specification of the optional parameters.

1
Purpose

nag_glopt_bnd_pso (e05sac) is designed to search for the global minimum or maximum of an arbitrary function, using Particle Swarm Optimization (PSO). Derivatives are not required, although these may be used by an accompanying local minimization function if desired. nag_glopt_bnd_pso (e05sac) is essentially identical to nag_glopt_nlp_pso (e05sbc), but with a simpler interface and with various optional parameters removed; otherwise most arguments are identical. In particular, nag_glopt_bnd_pso (e05sac) does not handle general constraints.

2
Specification

#include <nag.h>
#include <nage05.h>
void  nag_glopt_bnd_pso (Integer ndim, Integer npar, double xb[], double *fb, const double bl[], const double bu[],
void (*objfun)(Integer *mode, Integer ndim, const double x[], double *objf, double vecout[], Integer nstate, Nag_Comm *comm),
void (*monmod)(Integer ndim, Integer npar, double x[], const double xb[], double fb, const double xbest[], const double fbest[], const Integer itt[], Nag_Comm *comm, Integer *inform),
Integer iopts[], double opts[], Nag_Comm *comm, Integer itt[], Integer *inform, NagError *fail)
Before calling nag_glopt_bnd_pso (e05sac), nag_glopt_opt_set (e05zkc) must be called with optstr set to ‘Initialize = e05sac’. Optional parameters may also be specified by calling nag_glopt_opt_set (e05zkc) before the call to nag_glopt_bnd_pso (e05sac).

3
Description

nag_glopt_bnd_pso (e05sac) uses a stochastic method based on Particle Swarm Optimization (PSO) to search for the global optimum of a nonlinear function F, subject to a set of bound constraints on the variables. In the PSO algorithm (see Section 11), a set of particles is generated in the search space, and advances each iteration to (hopefully) better positions using a heuristic velocity based upon inertia, cognitive memory and global memory. The inertia is provided by a decreasingly weighted contribution from a particles current velocity, the cognitive memory refers to the best candidate found by an individual particle and the global memory refers to the best candidate found by all the particles. This allows for a global search of the domain in question.
Further, this may be coupled with a selection of local minimization functions, which may be called during the iterations of the heuristic algorithm, the interior phase, to hasten the discovery of locally optimal points, and after the heuristic phase has completed to attempt to refine the final solution, the exterior phase. Different options may be set for the local optimizer in each phase.
Without loss of generality, the problem is assumed to be stated in the following form:
minimize xRndim Fx   subject to   x u ,  
where the objective Fx is a scalar function, x is a vector in Rndim and the vectors u are lower and upper bounds respectively for the ndim variables. The objective function may be nonlinear. Continuity of F is not essential. For functions which are smooth and primarily unimodal, faster solutions will almost certainly be achieved by using Chapter e04 functions directly.
For functions which are smooth and multi-modal, gradient dependent local minimization functions may be coupled with nag_glopt_bnd_pso (e05sac).
For multi-modal functions for which derivatives cannot be provided, particularly functions with a significant level of noise in their evaluation, nag_glopt_bnd_pso (e05sac) should be used either alone, or coupled with nag_opt_simplex_easy (e04cbc).
The ndim lower and upper box bounds on the variable x are included to initialize the particle swarm into a finite hypervolume, although their subsequent influence on the algorithm is user determinable (see the option Boundary in Section 12). It is strongly recommended that sensible bounds are provided for all variables.
nag_glopt_bnd_pso (e05sac) may also be used to maximize the objective function (see the option Optimize).
Due to the nature of global optimization, unless a predefined target is provided, there is no definitive way of knowing when to end a computation. As such several stopping heuristics have been implemented into the algorithm. If any of these is achieved, nag_glopt_bnd_pso (e05sac) will exit with fail.code= NW_SOLUTION_NOT_GUARANTEED, and the parameter inform will indicate which criteria was reached. See inform for more information.
In addition, you may provide your own stopping criteria through monmod and objfun.
nag_glopt_nlp_pso (e05sbc) provides a comprehensive interface, allowing for the inclusion of general nonlinear constraints.

4
References

Gill P E, Murray W and Wright M H (1981) Practical Optimization Academic Press
Kennedy J and Eberhart R C (1995) Particle Swarm Optimization Proceedings of the 1995 IEEE International Conference on Neural Networks 1942–1948
Koh B, George A D, Haftka R T and Fregly B J (2006) Parallel Asynchronous Particle Swarm Optimization International Journal for Numerical Methods in Engineering 67(4) 578–595
Vaz A I and Vicente L N (2007) A Particle Swarm Pattern Search Method for Bound Constrained Global Optimization Journal of Global Optimization 39(2) 197–219 Kluwer Academic Publishers

5
Arguments

Note: for descriptions of the symbolic variables, see Section 11.
1:     ndim IntegerInput
On entry: ndim, the number of dimensions.
Constraint: ndim1.
2:     npar IntegerInput
On entry: npar, the number of particles to be used in the swarm. Assuming all particles remain within bounds, each complete iteration will perform at least npar function evaluations. Otherwise, significantly fewer objective function evaluations may be performed.
Suggested value: npar=10×ndim.
Constraint: npar5.
3:     xb[ndim] doubleOutput
On exit: the location of the best solution found, x~, in Rndim.
4:     fb double *Output
On exit: the objective value of the best solution, f~=Fx~.
5:     bl[ndim] const doubleInput
6:     bu[ndim] const doubleInput
On entry: bl is , the array of lower bounds, bu is u, the array of upper bounds. The ndim entries in bl and bu must contain the lower and upper simple (box) bounds of the variables respectively. These must be provided to initialize the sample population into a finite hypervolume, although their subsequent influence on the algorithm is user determinable (see the option Boundary in Section 12).
If bl[i-1]=bu[i-1] for any i1,,ndim, variable i will remain locked to bl[i-1] regardless of the Boundary option selected.
It is strongly advised that you place sensible lower and upper bounds on all variables, even if your model allows for variables to be unbounded (using the option Boundary=ignore) since these define the initial search space.
Constraints:
  • bl[i-1]bu[i-1], for i=1,2,,ndim;
  • bl[i-1]bu[i-1] for at least one i1,,ndim.
7:     objfun function, supplied by the userExternal Function
objfun must, depending on the value of mode, calculate the objective function and/or calculate the gradient of the objective function for a ndim-variable vector x. Gradients are only required if a local minimizer has been chosen which requires gradients. See the option Local Minimizer for more information.
The specification of objfun is:
void  objfun (Integer *mode, Integer ndim, const double x[], double *objf, double vecout[], Integer nstate, Nag_Comm *comm)
1:     mode Integer *Input/Output
On entry: indicates which functionality is required.
mode=0
Fx should be returned in objf. The value of objf on entry may be used as an upper bound for the calculation. Any expected value of Fx that is greater than objf may be approximated by this upper bound; that is objf can remain unaltered.
mode=1
Local Minimizer=e04ucc only
First derivatives can be evaluated and returned in vecout. Any unaltered elements of vecout will be approximated using finite differences.
mode=2
Local Minimizer=e04ucc only
Fx must be calculated and returned in objf, and available first derivatives can be evaluated and returned in vecout. Any unaltered elements of vecout will be approximated using finite differences.
mode=5
Fx must be calculated and returned in objf. The value of objf on entry may not be used as an upper bound.
mode=6
Local Minimizer=e04dgc only
All first derivatives must be evaluated and returned in vecout.
mode=7
Local Minimizer=e04dgc only
Fx must be calculated and returned in objf, and all first derivatives must be evaluated and returned in vecout.
On exit: if the value of mode is set to be negative, nag_glopt_bnd_pso (e05sac) will exit as soon as possible with fail.code= NE_USER_STOP and inform=mode.
2:     ndim IntegerInput
On entry: the number of dimensions.
3:     x[ndim] const doubleInput
On entry: x, the point at which the objective function and/or its gradient are to be evaluated.
4:     objf double *Input/Output
On entry: the value of objf passed to objfun varies with the argument mode.
mode=0
objf is an upper bound for the value of Fx, often equal to the best value of Fx found so far by a given particle. Only objective function values less than the value of objf on entry will be used further. As such this upper bound may be used to stop further evaluation when this will only increase the objective function value above the upper bound.
mode=1, 2, 5, 6 or 7
objf is meaningless on entry.
On exit: the value of objf returned varies with the argument mode.
mode=0
objf must be the value of Fx. Only values of Fx strictly less than objf on entry need be accurate.
mode=1 or 6
Need not be set.
mode=2, 5 or 7
Fx must be calculated and returned in objf. The entry value of objf may not be used as an upper bound.
5:     vecout[ndim] doubleInput/Output
On entry: if Local Minimizer=e04ucc, the values of vecout are used internally to indicate whether a finite difference approximation is required. See nag_opt_nlp (e04ucc).
On exit: the required values of vecout returned to the calling function depend on the value of mode.
mode=0 or 5
The value of vecout need not be set.
mode=1 or 2
vecout can contain components of the gradient of the objective function F xi  for some i=1,2,ndim, or acceptable approximations. Any unaltered elements of vecout will be approximated using finite differences.
mode=6 or 7
vecout must contain the gradient of the objective function F xi  for all i=1,2,ndim. Approximation of the gradient is strongly discouraged, and no finite difference approximations will be performed internally (see nag_opt_conj_grad (e04dgc)).
6:     nstate IntegerInput
On entry: nstate indicates various stages of initialization throughout the function. This allows for permanent global arguments to be initialized the least number of times. For example, you may initialize a random number generator seed.
nstate=2
objfun is called for the very first time. You may save computational time if certain data must be read or calculated only once.
nstate=1
objfun is called for the first time by a NAG local minimization function. You may save computational time if certain data required for the local minimizer need only be calculated at the initial point of the local minimization.
nstate=0
Used in all other cases.
7:     comm Nag_Comm *
Pointer to structure of type Nag_Comm; the following members are relevant to objfun.
userdouble *
iuserInteger *
pPointer 
The type Pointer will be void *. Before calling nag_glopt_bnd_pso (e05sac) you may allocate memory and initialize these pointers with various quantities for use by objfun when called from nag_glopt_bnd_pso (e05sac) (see Section 3.3.1.1 in How to Use the NAG Library and its Documentation).
Note: objfun should not return floating-point NaN (Not a Number) or infinity values, since these are not handled by nag_glopt_bnd_pso (e05sac). If your code inadvertently does return any NaNs or infinities, nag_glopt_bnd_pso (e05sac) is likely to produce unexpected results.
8:     monmod function, supplied by the userExternal Function
A user-specified monitoring and modification function. monmod is called once every complete iteration after a finalization check. It may be used to modify the particle locations that will be evaluated at the next iteration. This permits the incorporation of algorithmic modifications such as including additional advection heuristics and genetic mutations. monmod is only called during the main loop of the algorithm, and as such will be unaware of any further improvement from the final local minimization. If no monitoring and/or modification is required, monmod may be NULLFN.
The specification of monmod is:
void  monmod (Integer ndim, Integer npar, double x[], const double xb[], double fb, const double xbest[], const double fbest[], const Integer itt[], Nag_Comm *comm, Integer *inform)
1:     ndim IntegerInput
On entry: the number of dimensions.
2:     npar IntegerInput
On entry: the number of particles.
3:     x[ndim×npar] doubleInput/Output
Note: the ith component of the jth particle, xji, is stored in x[j-1×ndim+i-1].
On entry: the npar particle locations, xj, which will currently be used during the next iteration unless altered in monmod.
On exit: the particle locations to be used during the next iteration.
4:     xb[ndim] const doubleInput
On entry: the location, x~, of the best solution yet found.
5:     fb doubleInput
On entry: the objective value, f~=Fx~, of the best solution yet found.
6:     xbest[ndim×npar] const doubleInput
Note: the ith component of the position of the jth particle's cognitive memory, x^ji, is stored in xbest[j-1×ndim+i-1].
On entry: the locations currently in the cognitive memory, x^j, for j=1,2,,npar (see Section 11).
7:     fbest[npar] const doubleInput
On entry: the objective values currently in the cognitive memory, Fx^j, for j=1,2,,npar.
8:     itt[6] const IntegerInput
On entry: iteration and function evaluation counters (see description of itt below).
9:     comm Nag_Comm *
Pointer to structure of type Nag_Comm; the following members are relevant to monmod.
userdouble *
iuserInteger *
pPointer 
The type Pointer will be void *. Before calling nag_glopt_bnd_pso (e05sac) you may allocate memory and initialize these pointers with various quantities for use by monmod when called from nag_glopt_bnd_pso (e05sac) (see Section 3.3.1.1 in How to Use the NAG Library and its Documentation).
10:   inform Integer *Input/Output
On entry: inform=0
On exit: setting inform<0 will cause near immediate exit from nag_glopt_bnd_pso (e05sac). This value will be returned as inform with fail.code= NE_USER_STOP. You need not set inform unless you wish to force an exit.
Note: monmod should not return floating-point NaN (Not a Number) or infinity values, since these are not handled by nag_glopt_bnd_pso (e05sac). If your code inadvertently does return any NaNs or infinities, nag_glopt_bnd_pso (e05sac) is likely to produce unexpected results.
9:     iopts[dim] IntegerCommunication Array
Note: the dimension, dim, of this array is dictated by the requirements of associated functions that must have been previously called. This array MUST be the same array passed as argument iopts in the previous call to nag_glopt_opt_set (e05zkc).
On entry: optional parameter array as generated and possibly modified by calls to nag_glopt_opt_set (e05zkc). The contents of iopts MUST NOT be modified directly between calls to nag_glopt_bnd_pso (e05sac), nag_glopt_opt_set (e05zkc) or nag_glopt_opt_get (e05zlc).
10:   opts[dim] doubleCommunication Array
Note: the dimension, dim, of this array is dictated by the requirements of associated functions that must have been previously called. This array MUST be the same array passed as argument opts in the previous call to nag_glopt_opt_set (e05zkc).
On entry: optional parameter array as generated and possibly modified by calls to nag_glopt_opt_set (e05zkc). The contents of opts MUST NOT be modified directly between calls to nag_glopt_bnd_pso (e05sac), nag_glopt_opt_set (e05zkc) or nag_glopt_opt_get (e05zlc).
11:   comm Nag_Comm *
The NAG communication argument (see Section 3.3.1.1 in How to Use the NAG Library and its Documentation).
12:   itt[6] IntegerOutput
On exit: integer iteration counters for nag_glopt_bnd_pso (e05sac).
itt[0]
Number of complete iterations.
itt[1]
Number of complete iterations without improvement to the current optimum.
itt[2]
Number of particles converged to the current optimum.
itt[3]
Number of improvements to the optimum.
itt[4]
Number of function evaluations performed.
itt[5]
Number of particles reset.
13:   inform Integer *Output
On exit: indicates which finalization criterion was reached. The possible values of inform are:
inform Meaning
<0 Exit from a user-supplied function.
0 nag_glopt_bnd_pso (e05sac) has detected an error and terminated.
1 The provided objective target has been achieved. (Target Objective Value).
2 The standard deviation of the location of all the particles is below the set threshold (Swarm Standard Deviation). If the solution returned is not satisfactory, you may try setting a smaller value of Swarm Standard Deviation, or try adjusting the options governing the repulsive phase (Repulsion Initialize, Repulsion Finalize).
3 The total number of particles converged (Maximum Particles Converged) to the current global optimum has reached the set limit. This is the number of particles which have moved to a distance less than Distance Tolerance from the optimum with regard to the L2 norm. If the solution is not satisfactory, you may consider lowering the Distance Tolerance. However, this may hinder the global search capability of the algorithm.
4 The maximum number of iterations without improvement (Maximum Iterations Static) has been reached, and the required number of particles (Maximum Iterations Static Particles) have converged to the current optimum. Increasing either of these options will allow the algorithm to continue searching for longer. Alternatively if the solution is not satisfactory, re-starting the application several times with Repeatability=OFF may lead to an improved solution.
5 The maximum number of iterations (Maximum Iterations Completed) has been reached. If the number of iterations since improvement is small, then a better solution may be found by increasing this limit, or by using the option Local Minimizer with corresponding exterior options. Otherwise if the solution is not satisfactory, you may try re-running the application several times with Repeatability=OFF and a lower iteration limit, or adjusting the options governing the repulsive phase (Repulsion Initialize, Repulsion Finalize).
6 The maximum allowed number of function evaluations (Maximum Function Evaluations) has been reached. As with inform=5, increasing this limit if the number of iterations without improvement is small, or decreasing this limit and running the algorithm multiple times with Repeatability=OFF, may provide a superior result.
14:   fail NagError *Input/Output
The NAG error argument (see Section 3.7 in How to Use the NAG Library and its Documentation).
nag_glopt_bnd_pso (e05sac) will return fail.code= NE_NOERROR if and only if a finalization criterion has been reached which can guarantee success. This may only happen if the option Target Objective Value has been set and reached at a point within the search domain. The finalization criterion Target Objective Value is not activated using default option settings, and must be explicitly set using nag_glopt_opt_set (e05zkc) if required.
nag_glopt_bnd_pso (e05sac) will return fail.code= NW_SOLUTION_NOT_GUARANTEED if no error has been detected, and a finalization criterion has been achieved which cannot guarantee success. This does not indicate that the function has failed, merely that the returned solution cannot be guaranteed to be the true global optimum.
The value of inform should be examined to determine which finalization criterion was reached.

6
Error Indicators and Warnings

NE_ALLOC_FAIL
Dynamic memory allocation failed.
See Section 2.3.1.2 in How to Use the NAG Library and its Documentation for further information.
NE_BAD_PARAM
On entry, argument value had an illegal value.
NE_BOUND
On entry, bl[i]=bu[i] for all i.
Constraint: bu[i]>bl[i] for at least one i.
On entry, bl[value]=value and bu[value]=value.
Constraint: bu[i]bl[i] for all i.
NE_DERIV_ERRORS
Derivative checks indicate possible errors in the supplied derivatives. Gradient checks may be disabled by setting Verify Gradients=OFF.
NE_INT
On entry, ndim=value.
Constraint: ndim1.
On entry, npar=value.
Constraint: npar5×num_threads, where num_threads is the value returned by the OpenMP environment variable OMP_NUM_THREADS, or num_threads is 1 for a serial version of this function.
NE_INTERNAL_ERROR
An internal error has occurred in this function. Check the function call and any array sizes. If the call is correct then please contact NAG for assistance.
See Section 2.7.6 in How to Use the NAG Library and its Documentation for further information.
NE_INVALID_OPTION
Either the option arrays have not been initialized for nag_glopt_bnd_pso (e05sac), or they have become corrupted.
NE_NO_LICENCE
Your licence key may have expired or may not have been installed correctly.
See Section 2.7.5 in How to Use the NAG Library and its Documentation for further information.
NE_USER_STOP
User requested exit value during call to monmod.
User requested exit value during call to objfun.
NW_FAST_SOLUTION
If the option Target Warning has been activated, this indicates that the Target Objective Value has been achieved to specified tolerances at a sufficiently constrained point, either during the initialization phase, or during the first two iterations of the algorithm. While this is not necessarily an error, it may occur if:
(i) The target was achieved at the first point sampled by the function. This will be the mean of the lower and upper bounds.
(ii) The target may have been achieved at a randomly generated sample point. This will always be a possibility provided that the domain under investigation contains a point with a target objective value.
(iii) If the Local Minimizer has been set, then a sample point may have been inside the basin of attraction of a satisfactory point. If this occurs repeatedly when the function is called, it may imply that the objective is largely unimodal, and that it may be more efficient to use the function selected as the Local Minimizer directly.
Assuming that objfun is correct, you may wish to set a better Target Objective Value, or a stricter Target Objective Tolerance.
NW_SOLUTION_NOT_GUARANTEED
A finalization criterion was reached that cannot guarantee success.
On exit, inform=value.

7
Accuracy

If fail.code= NE_NOERROR (or fail.code= NW_FAST_SOLUTION) or fail.code= NW_SOLUTION_NOT_GUARANTEED on exit, either a Target Objective Value or finalization criterion has been reached, depending on user selected options. As with all global optimization software, the solution achieved may not be the true global optimum. Various options allow for either greater search diversity or faster convergence to a (local) optimum (See Sections 11 and 12).
Provided the objective function and constraints are sufficiently well behaved, if a local minimizer is used in conjunction with nag_glopt_bnd_pso (e05sac), then it is more likely that the final result will at least be in the near vicinity of a local optimum, and due to the global search characteristics of the particle swarm, this solution should be superior to many other local optima.
Caution should be used in accelerating the rate of convergence, as with faster convergence, less of the domain will remain searchable by the swarm, making it increasingly difficult for the algorithm to detect the basin of attraction of superior local optima. Using the options Repulsion Initialize and Repulsion Finalize described in Section 12 will help to overcome this, by causing the swarm to diverge away from the current optimum once no more local improvement is likely.
On successful exit with guaranteed success, fail.code= NE_NOERROR. This may only happen if a Target Objective Value is assigned and is reached by the algorithm.
On successful exit without guaranteed success, fail.code= NW_SOLUTION_NOT_GUARANTEED is returned. This will happen if another finalization criterion is achieved without the detection of an error.
In both cases, the value of inform provides further information as to the cause of the exit.

8
Parallelism and Performance

The code for nag_glopt_bnd_pso (e05sac) is not directly threaded for parallel execution. In particular, none of the user-supplied functions will be called from within a parallel region generated by nag_glopt_bnd_pso (e05sac).

9
Further Comments

The memory used by nag_glopt_bnd_pso (e05sac) is relatively static throughout. As such, nag_glopt_bnd_pso (e05sac) may be used in problems with high dimension number (ndim>100) without the concern of computational resource exhaustion, although the probability of successfully locating the global optimum will decrease dramatically with the increase in dimensionality.
Due to the stochastic nature of the algorithm, the result will vary over multiple runs. This is particularly true if arguments and options are chosen to accelerate convergence at the expense of the global search. However, the option Repeatability=ON may be set to initialize the internal random number generator using a preset seed, which will result in identical solutions being obtained.

10
Example

This example uses a particle swarm to find the global minimum of the Schwefel function:
minimize xRndim f = i=1 ndim xi sinxi  
xi -500,500 ,  for ​ i=1,2,,ndim .  
In two dimensions the optimum is fmin=-837.966, located at x=-420.97,-420.97.
The example demonstrates how to initialize and set the options arrays using nag_glopt_opt_set (e05zkc), how to query options using nag_glopt_opt_get (e05zlc), and finally how to search for the global optimum using nag_glopt_bnd_pso (e05sac). The function is minimized several times to demonstrate using nag_glopt_bnd_pso (e05sac) alone, and coupled with local minimizers. This program uses the non-default option Repeatability=ON to produce repeatable solutions.

10.1
Program Text

Program Text (e05sace.c)

10.2
Program Data

None.

10.3
Program Results

Program Results (e05sace.r)

11
Algorithmic Details

The following pseudo-code describes the algorithm used with the repulsion mechanism.
INITIALIZE for ​ j=1, ​ npar xj = R Ubox,ubox x^ j = R Ubox,ubox vj = R U -V max ,Vmax f^j = F x^ j initialize ​wj wj = Wmax Weight Initialize=MAXIMUM Wini Weight Initialize=INITIAL R U W min , W max Weight Initialize=RANDOMIZED end for x~ = 12 box + ubox f~ = F x~ Ic = Is = 0 SWARM while (not finalized), Ic = Ic + 1 for ​ j = 1 , npar xj = BOUNDARYxj,box,ubox fj = F xj if ​ fj < f^j f^j = fj , ​ x^j = xj if ​ fj < f~ f~ = fj , ​ x~ = xj end for if ​ new f~ LOCMINx~,f~,Oi , ​ Is=0 [see note on repulsion below for code insertion] else Is = Is + 1 for ​ j = 1 , npar vj = wj vj + Cs D1 x^j - xj + Cg D2 x~ - xj xj = xj + vj if ​ xj - x~ < dtol reset ​ xj, ​ vj, ​ wj; ​ x^j = xj else update ​wj end for if (target achieved or termination criterion satisfied) finalized=true monmod xj end LOCMINx~,f~,Oe  
The definition of terms used in the above pseudo-code are as follows.
npar  the number of particles, npar
box array of ndim lower box bounds
ubox array of ndim upper box bounds
xj position of particle j
x^j best position found by particle j
x~ best position found by any particle
fj Fxj
f^j Fx^j, best value found by particle j
f~ Fx~, best value found by any particle
vj velocity of particle j
wj weight on vj for velocity update, decreasing according to Weight Decrease
Vmax maximum absolute velocity, dependent upon Maximum Variable Velocity
Ic swarm iteration counter
Is iterations since x~ was updated
D1,D2 diagonal matrices with random elements in range 0,1
Cs the cognitive advance coefficient which weights velocity towards x^j, adjusted using Advance Cognitive
Cg the global advance coefficient which weights velocity towards x~, adjusted using Advance Global
dtol the Distance Tolerance for resetting a converged particle
RUbox,ubox an array of random numbers whose ith element is drawn from a uniform distribution in the range boxi,uboxi, for i=1,2,,ndim
Oi local optimizer interior options
Oe local optimizer exterior options
LOCMINx,f,O apply local optimizer using the set of options O using the solution x,f as the starting point, if used (not default)
monmod monitor progress and possibly modify xj
BOUNDARY apply required behaviour for xj outside bounding box, (see Boundary)
new (f~) true if x~, c~, f~ were updated at this iteration
Additionally a repulsion phase can be introduced by changing from the default values of options Repulsion Finalize (rf), Repulsion Initialize (ri) and Repulsion Particles (rp). If the number of static particles is denoted ns then the following can be inserted after the new(f~) check in the pseudo-code above.
else ​ if ​ ( ns rp ​ and ​ ri Is ri + rf ) LOCMINx~,f~,Oi use ​ -Cg ​ instead of ​ Cg ​ in velocity updates if ​ Is = ri + rf Is = 0  

12
Optional Parameters

This section can be skipped if you wish to use the default values for all optional parameters, otherwise, the following is a list of the optional parameters available and a full description of each optional parameter is provided in Section 12.1.

12.1
Description of the Optional Parameters

For each option, we give a summary line, a description of the optional parameter and details of constraints.
The summary line contains:
All options accept the value ‘DEFAULT’ in order to return single options to their default states.
Keywords and character values are case insensitive, however they must be separated by at least one space.
For nag_glopt_bnd_pso (e05sac) the maximum length of the argument cvalue used by nag_glopt_opt_get (e05zlc) is 15.
Advance CognitiverDefault=2.0
The cognitive advance coefficient, Cs. When larger than the global advance coefficient, this will cause particles to be attracted toward their previous best positions. Setting r=0.0 will cause nag_glopt_bnd_pso (e05sac) to act predominantly as a local optimizer. Setting r>2.0 may cause the swarm to diverge, and is generally inadvisable. At least one of the global and cognitive coefficients must be nonzero.
Advance GlobalrDefault=2.0
The global advance coefficient, Cg. When larger than the cognitive coefficient this will encourage convergence toward the best solution yet found. Values r0,1 will inhibit particles overshooting the optimum. Values r1,2 cause particles to fly over the optimum some of the time. Larger values can prohibit convergence. Setting r=0.0 will remove any attraction to the current optimum, effectively generating a Monte–Carlo multi-start optimization algorithm. At least one of the global and cognitive coefficients must be nonzero.
BoundaryaDefault=FLOATING
Determines the behaviour if particles leave the domain described by the box bounds. This only affects the general PSO algorithm, and will not pass down to any NAG local minimizers chosen.
This option is only effective in those dimensions for which bl[i-1]bu[i-1], i=1,2,,ndim.
IGNORE
The box bounds are ignored. The objective function is still evaluated at the new particle position.
RESET
The particle is re-initialized inside the domain. x^j and f^j are not affected.
FLOATING
The particle position remains the same, however the objective function will not be evaluated at the next iteration. The particle will probably be advected back into the domain at the next advance due to attraction by the cognitive and global memory.
HYPERSPHERICAL
The box bounds are wrapped around an ndim-dimensional hypersphere. As such a particle leaving through a lower bound will immediately re-enter through the corresponding upper bound and vice versa. The standard distance between particles is also modified accordingly.
FIXED
The particle rests on the boundary, with the corresponding dimensional velocity set to 0.0.
Distance ScalingaDefault=ON
Determines whether distances should be scaled by box widths.
ON
When a distance is calculated between x and y, a scaled L2 norm is used.
L2 x,y = i | ui i , indim xi - yi ui - i 2 1 2 .  
OFF
Distances are calculated as the standard L2 norm without any rescaling.
L2 x,y = i=1 ndim xi - yi 2 1 2 .  
Distance TolerancerDefault=10-4
This is the distance, dtol between particles and the global optimum which must be reached for the particle to be considered converged, i.e., that any subsequent movement of such a particle cannot significantly alter the global optimum. Once achieved the particle is reset into the box bounds to continue searching.
Constraint: r>0.0.
Function PrecisionrDefault=ε0.9
The parameter defines εr, which is intended to be a measure of the accuracy with which the problem function Fx can be computed. If r<ε or r1, the default value is used.
The value of εr should reflect the relative precision of 1+Fx; i.e., εr acts as a relative precision when F is large, and as an absolute precision when F is small. For example, if Fx is typically of order 1000 and the first six significant digits are known to be correct, an appropriate value for εr would be 10-6. In contrast, if Fx is typically of order 10-4 and the first six significant digits are known to be correct, an appropriate value for εr would be 10-10. The choice of εr can be quite complicated for badly scaled problems; see Chapter 8 of Gill et al. (1981) for a discussion of scaling techniques. The default value is appropriate for most simple functions that are computed with full accuracy. However when the accuracy of the computed function values is known to be significantly worse than full precision, the value of εr should be large enough so that no attempt will be made to distinguish between function values that differ by less than the error inherent in the calculation.
Local Boundary RestrictionrDefault=0.5
Contracts the box boundaries used by a box constrained local minimizer to, βl,βu, containing the start point x, where
i = r × ui - i βli = maxi, xi - i2 βui = minui, xi + i2 ,  i=1,,ndim .  
Smaller values of r thereby restrict the size of the domain exposed to the local minimizer, possibly reducing the amount of work done by the local minimizer.
Constraint: 0.0r1.0.
Local Interior Iterationsi1
Local Interior Major Iterationsi1
Local Exterior Iterationsi2
Local Exterior Major Iterationsi2
The maximum number of iterations or function evaluations the chosen local minimizer will perform inside (outside) the main loop if applicable. For the NAG minimizers these correspond to:
Minimizer Parameter/option Default Interior Default Exterior
nag_opt_simplex_easy (e04cbc) maxcal ndim+10 2×ndim+15
nag_opt_conj_grad (e04dgc) Iteration Limit max30,3×ndim max50,5×ndim
nag_opt_nlp (e04ucc) Major Iteration Limit max10,2×ndim max30,3×ndim
Unless set, these are functions of the parameters passed to nag_glopt_bnd_pso (e05sac).
Setting i=0 will disable the local minimizer in the corresponding algorithmic region. For example, setting Local Interior Iterations=0 and Local Exterior Iterations=30 will cause the algorithm to perform no local minimizations inside the main loop of the algorithm, and a local minimization with upto 30 iterations after the main loop has been exited.
Constraint: i10i20.
Local Interior Tolerancer1Default=10-4
Local Exterior Tolerancer2Default=10-4
This is the tolerance provided to a local minimizer in the interior (exterior) of the main loop of the algorithm.
Constraint: r1>0.0,r2>0.0.
Local Interior Minor Iterationsi1
Local Exterior Minor Iterationsi2
Where applicable, the secondary number of iterations the chosen local minimizer will use inside (outside) the main loop. Currently the relevant default values are:
Minimizer Parameter/option Default Interior Default Exterior
nag_opt_nlp (e04ucc) Minor Iteration Limit max10,2×ndim max30,3×ndim
Constraint: i10i20.
Local MinimizeraDefault=OFF
Allows for a choice of Chapter e04 functions to be used as a coupled, dedicated local minimizer.
OFF
No local minimization will be performed in either the INTERIOR or EXTERIOR sections of the algorithm.
e04cbc
Use nag_opt_simplex_easy (e04cbc) as the local minimizer. This does not require the calculation of derivatives.
On a call to objfun during a local minimization, mode=5.
e04dgc
Use nag_opt_conj_grad (e04dgc) as the local minimizer.
Accurate derivatives must be provided, and will not be approximated internally. Additionally, each call to objfun during a local minimization will require either the objective to be evaluated alone, or both the objective and its gradient to be evaluated. Hence on a call to objfun, mode=5 or 7.
e04ucc
Use nag_opt_nlp (e04ucc) as the local minimizer. This operates such that any derivatives of the objective function that you cannot supply, will be approximated internally using finite differences.
Either, the objective, objective gradient, or both may be requested during a local minimization, and as such on a call to objfun, mode=1, 2 or 5.
The box bounds forwarded to this function from nag_glopt_bnd_pso (e05sac) will have been acted upon by Local Boundary Restriction. As such, the domain exposed may be greatly smaller than that provided to nag_glopt_bnd_pso (e05sac).
Maximum Function EvaluationsiDefault =Imax
The maximum number of evaluations of the objective function. When reached this will return fail.code= NW_SOLUTION_NOT_GUARANTEED and inform=6.
Constraint: i>0.
Maximum Iterations CompletediDefault=1000×ndim
The maximum number of complete iterations that may be performed. Once exceeded nag_glopt_bnd_pso (e05sac) will exit with fail.code= NW_SOLUTION_NOT_GUARANTEED and inform=5.
Unless set, this adapts to the parameters passed to nag_glopt_bnd_pso (e05sac).
Constraint: i1.
Maximum Iterations StaticiDefault=100
The maximum number of iterations without any improvement to the current global optimum. If exceeded nag_glopt_bnd_pso (e05sac) will exit with fail.code= NW_SOLUTION_NOT_GUARANTEED and inform=4. This exit will be hindered by setting Maximum Iterations Static Particles to larger values.
Constraint: i1.
Maximum Iterations Static ParticlesiDefault=0
The minimum number of particles that must have converged to the current optimum before the function may exit due to Maximum Iterations Static with fail.code= NW_SOLUTION_NOT_GUARANTEED and inform=4.
Constraint: i0.
Maximum Particles ConvergediDefault =Imax
The maximum number of particles that may converge to the current optimum. When achieved, nag_glopt_bnd_pso (e05sac) will exit with fail.code= NW_SOLUTION_NOT_GUARANTEED and inform=3. This exit will be hindered by setting ‘Repulsion’ options, as these cause the swarm to re-expand.
Constraint: i>0.
Maximum Particles ResetiDefault =Imax
The maximum number of particles that may be reset after converging to the current optimum. Once achieved no further particles will be reset, and any particles within Distance Tolerance of the global optimum will continue to evolve as normal.
Constraint: i>0.
Maximum Variable VelocityrDefault=0.25
Along any dimension j, the absolute velocity is bounded above by vjr×uj-j=Vmax. Very low values will greatly increase convergence time. There is no upper limit, although larger values will allow more particles to be advected out of the box bounds, and values greater than 4.0 may cause significant and potentially unrecoverable swarm divergence.
Constraint: r>0.0.
OptimizeaDefault=MINIMIZE
Determines whether to maximize or minimize the objective function.
MINIMIZE
The objective function will be minimized.
MAXIMIZE
The objective function will be maximized. This is accomplished by minimizing the negative of the objective.
RepeatabilityaDefault=OFF
Allows for the same random number generator seed to be used for every call to nag_glopt_bnd_pso (e05sac). Repeatability=OFF is recommended in general.
OFF
The internal generation of random numbers will be nonrepeatable.
ON
The same seed will be used.
Repulsion FinalizeiDefault =Imax
The number of iterations performed in a repulsive phase before re-contraction. This allows a re-diversified swarm to contract back toward the current optimum, allowing for a finer search of the near optimum space.
Constraint: i2.
Repulsion InitializeiDefault =Imax
The number of iterations without any improvement to the global optimum before the algorithm begins a repulsive phase. This phase allows the particle swarm to re-expand away from the current optimum, allowing more of the domain to be investigated. The repulsive phase is automatically ended if a superior optimum is found.
Constraint: i2.
Repulsion ParticlesiDefault=0
The number of particles required to have converged to the current optimum before any repulsive phase may be initialized. This will prevent repulsion before a satisfactory search of the near optimum area has been performed, which may happen for large dimensional problems.
Constraint: i0.
SeediDefault=0
Sets the random number generator seed to be used when Repeatability=ON. If set to 0, the default seed will be used. If not, the absolute value of Seed will be used to generate the random number generator seed.
Swarm Standard DeviationrDefault=0.1
The target standard deviation of the particle distances from the current optimum. Once the standard deviation is below this level, nag_glopt_bnd_pso (e05sac) will exit with fail.code= NW_SOLUTION_NOT_GUARANTEED and inform=2. This criterion will be penalized by the use of ‘Repulsion’ options, as these cause the swarm to re-expand, increasing the standard deviation of the particle distances from the best point.
Constraint: r0.0.
Target ObjectiveaDefault=OFF
Target Objective ValuerDefault=0.0
Activate or deactivate the use of a target value as a finalization criterion. If active, then once the supplied target value for the objective function is found (beyond the first iteration if Target Warning is active) nag_glopt_bnd_pso (e05sac) will exit with fail.code= NE_NOERROR and inform=1. Other than checking for feasibility only (Optimize=CONSTRAINTS), this is the only finalization criterion that guarantees that the algorithm has been successful. If the target value was achieved at the initialization phase or first iteration and Target Warning is active, nag_glopt_bnd_pso (e05sac) will exit with fail.code= NW_FAST_SOLUTION. This option may take any real value r, or the character ON/OFF as well as DEFAULT. If this option is queried using nag_glopt_opt_get (e05zlc), the current value of r will be returned in rvalue, and cvalue will indicate whether this option is ON or OFF. The behaviour of the option is as follows:
r
Once a point is found with an objective value within the Target Objective Tolerance of r, nag_glopt_bnd_pso (e05sac) will exit successfully with fail.code= NE_NOERROR and inform=1.
OFF
The current value of r will remain stored, however it will not be used as a finalization criterion.
ON
The current value of r stored will be used as a finalization criterion.
DEFAULT
The stored value of r will be reset to its default value (0.0), and this finalization criterion will be deactivated.
Target Objective SafeguardrDefault=100.0ε
If you have given a target objective value to reach in objval (the value of the optional parameter Target Objective Value), objsfg sets your desired safeguarded termination tolerance, for when objval is close to zero.
Constraint: objsfg2ε.
Target Objective TolerancerDefault=0.0
The optional tolerance to a user-specified target value.
Constraint: r0.0.
Target WarningaDefault=OFF
Activates or deactivates the error exit associated with the target value being achieved before entry into the main loop of the algorithm, fail.code= NW_FAST_SOLUTION.
OFF
No error will be returned, and the function will exit normally.
ON
An error will be returned if the target objective is reached prematurely, and the function will exit with fail.code= NW_FAST_SOLUTION.
Verify GradientsaDefault=ON
Adjusts the level of gradient checking performed when gradients are required. Gradient checks are only performed on the first call to the chosen local minimizer if it requires gradients. There is no guarantee that the gradient check will be correct, as the finite differences used in the gradient check are themselves subject to inaccuracies.
OFF
No gradient checking will be performed.
ON
A cheap gradient check will be performed on both the gradients corresponding to the objective through objfun.
OBJECTIVE
FULL
A more expensive gradient check will be performed on the gradients corresponding to the objective objfun.
Weight DecreaseaDefault=INTEREST
Determines how particle weights decrease.
OFF
Weights do not decrease.
INTEREST
Weights decrease through compound interest as wIT+1=wIT1-Wval, where Wval is the Weight Value and IT is the current number of iterations.
LINEAR
Weights decrease linearly following wIT+1=wIT-IT×Wmax-Wmin/ITmax, where IT is the iteration number and ITmax is the maximum number of iterations as set by Maximum Iterations Completed.
Weight InitialrDefault=Wmax
The initial value of any particle's inertial weight, Wini, or the minimum possible initial value if initial weights are randomized. When set, this will override Weight Initialize=RANDOMIZED or MAXIMUM, and as such these must be set afterwards if so desired.
Constraint: WminrWmax.
Weight InitializeaDefault=MAXIMUM
Determines how the initial weights are distributed.
INITIAL
All weights are initialized at the initial weight, Wini, if set. If Weight Initial has not been set, this will be the maximum weight, Wmax.
MAXIMUM
All weights are initialized at the maximum weight, Wmax.
RANDOMIZED
Weights are uniformly distributed in Wmin,Wmax or Wini,Wmax if Weight Initial has been set.
Weight MaximumrDefault=1.0
The maximum particle weight, Wmax.
Constraint: 1.0rWmin (If Wini has been set then 1.0rWini.)
Weight MinimumrDefault=0.1
The minimum achievable weight of any particle, Wmin. Once achieved, no further weight reduction is possible.
Constraint: 0.0rWmax (If Wini has been set then 0.0rWini.)
Weight ResetaDefault=MAXIMUM
Determines how particle weights are re-initialized.
INITIAL
Weights are re-initialized at the initial weight if set. If Weight Initial has not been set, this will be the maximum weight.
MAXIMUM
Weights are re-initialized at the maximum weight.
RANDOMIZED
Weights are uniformly distributed in Wmin,Wmax or Wini,Wmax if Weight Initial has been set.
Weight ValuerDefault=0.01
The constant Wval used with Weight Decrease=INTEREST.
Constraint: 0.0r13.