LatNet Builder Manual
2.0.1-11
Software Package for Constructing Highly Uniform Point Sets
|
The latnetbuilder
executable supports the following command-line options:
--version
--help
/ -h
--set-type
/ -t
lattice
to search for rank-1 lattices; ornet
to search for digital nets. --construction
/ -c
Required. Specifies the construction type. Takes one of the following values:
–set-type lattice
):ordinary
to search for ordinary lattices; orpolynomial
to search for polynomial lattices.–set-type net
):sobol
to search for Sobol' nets; orpolynomial
to search for polynomial lattice rules; orexplicit
to search for digital nets with explicit generating matrices.You can learn more on the different point set types and constructions here. Details on how to use the two options above can be find here.
--dimension
/ -d
--size-parameter
/ -s
Required. Specifies the size parameter of the point sets. Takes an argument of the form:
–set-type lattice
):–size-parameter modulus
; or–size-parameter primeBase^power
; or–size-parameter 2^k
(even for polynomial lattice rules).–set-type net
):–size-parameter 2^k
; or–size-parameter modulus
(only for polynomial lattice rules); or–size-parameter primeBase^power
(only for polynomial lattice rules).See here for additional details.
--exploration-method
/ -e
Required. Specifies the exploration method. Takes an argument of the form:
evaluate:point-set-description
to compute the merit value of the point set described by point-set-description
. See here for details about point-set-description
.exhaustive
for exhaustive search;random:samples
for a random search with samples
random samples;full-CBC
for a component-by-component search;random-CBC:samples
for a random component-by-component search with samples
random samples per coordinate;
Specific to lattices (–set-type lattice
):
Korobov
for a Korobov search;random-Korobov:samples
for a random Korobov search with samples
random samples;fast-CBC
for a fast CBC search (requires a coordinate-uniform implementation of the selected figure of merit);
Recall that the implementation of the fast CBC algorithm only supports modulus that are a power of a prime base in the ordinary case and irreducible modulus in the polynomial case.
extend:modulus:genVec
to extend the lattice to a lattice with modulus modulus
and generating vector genVec
specified as a dash-separated list of integers/polynomials. See here for details about genVec
.
Specific to digital nets (–set-type net
):
mixed-CBC:samples:nbFull
for a full-CBC search for the first nbFull
coordinates and then a random-CBC search with samples
random samples for the remaining coordinates.When the random variant of a search is used with a filter (see the --filters
option below), the candidate samples that are rejected by the filter pipeline are not considered as valid samples, meaning that the user-specified number of random samples only corresponds to the number of accepted candidates.
--figure-of-merit
/ -f
Required. Specifies the figure of merit to be used. Takes an argument of either of the following forms:
figure
CU:figure
where the optional prefix CU:
indicates that a coordinate-uniform implementation of the evaluation algorithm should be used (only available for Palpha
, R1
, R
with –norm-type 2
and IAalpha
, IB
with –norm-type 1
)and where figure
is one of:
Palpha
for the weighted \(\mathcal P_\alpha\) discrepancy with \(\alpha=\)alpha
(the figures of merit are different for ordinary and polynomial lattice rules but have the same name);spectral
for the spectral figure of merit (only available with ordinary lattice rules);Ralpha
for the weighted \(\mathcal R_\alpha\) figure of merit with \(\alpha=\)alpha
(only available with ordinary lattice rules); orR
for the weighted \(\mathcal R\) figure of merit (only available with polynomial lattice rules and digital nets);t-value
for the t-value merit (only available with digital nets and incompatible with CBC explorations);projdep:t-value
for the projection-dependent t-value merit (only available wih digital nets);projdep:resolution-gap
for the projection-dependent resolution-gap (only available wih digital nets);IAalpha
for the interlaced \(B_{\alpha, d, (1)}\) discrepancy with \(\alpha=\)alpha
(only available for interlaced polynomial lattice rules and digital nets); orIB
for the interlaced \(B_{d, (2)}\) discrepancy (only available for interlaced polynomial lattice rules and digital nets).The definitions of all these figures of merit can be find here. More details on how to use this option are available on this page.
--interlacing-factor
/ -i
--norm-type
/ -q
Required. Specifies the type of the \(\ell_q\) norm to combine the merit values across projections. Takes an argument of the form:
realNumber
for a finite positive number;inf
for infinity. The value of this option must be compatible with the type of figure of merit, especially regarding the use of the coordinate-uniform evaluation algorithm.See Figures of merit for additional details on this parameter.
--weights
/ -w
Required. Specifies the type(s) of weights and their values. Takes a whitespace-separated list of arguments, each of which specifying a type of weights with its values. (The actual weights are the sum of these.) Takes argument of the form:
product:default:list
for product weights with the weights for the first coordinates specified by the comma-separated list of weights list
, and with weight default
for the other coordinates;order-dependent:default:list
for order-dependent weights with the weights for the first orders specified by the comma-separated list of weights list
(starting at order 1), and with weight default
for the other orders;POD:default1:list1:default2:list2
for POD weights, where default1
and list1
specify the default and individual order-dependent weights, and where default2
and list2
specify the default and individual product weights.projection-dependent:proj:weight:...:proj:weight
for projection-dependent weights where proj
is a projection (comma-separated list of coordinates) and weight
the associated weight. –weights file:path_to_file
to assign the weight weight
to all other where path_to_file
is the path to the file containing the weights. The file must consist of lines of the form:coordinates: weight
to assign the weight weight
to the projection defined by the comma-separated list of coordinates coordinates
;order order: weight
to assign the weight weight
to all other projections of order order
;default: weight
projections;#comment
to ignore comment
.The way to specify the different types of weights is explained here.
--weights-power
/ -p
Optional (default: same value as for the –norm-type
option). Specifies that the weights passed by option –weights
have already been elevated to that power.
By default, the weights are assumed to have been elevated to the same power as specified by option –norm-type
. A value of inf
is mapped to 1. See the definition of \(p\) in the first paragraph of this page for additional details.
--multilevel
/ -M
false
to search for unilevel point sets; ortrue
to search for multilevel point sets. --combiner
/ -C
sum
to sum the individual merit values of all nested levels;max
to select the maximum individual merit value across all nested levels;level:max
to select the individual merit value of the nested level with the largest number of points (the highest level);level:m
to select the individual merit value of the \(m\)th nested level. --filters
/ -F
Optional. Configures filters for merit values. Only supported when –set-type lattice
is set. Takes a whitespace-separated list of arguments of the form:
norm:type
for a normalizer of type specified by type
;low-pass:threshold
for a low-pass filter with threshold value threshold
.If the option –multilevel true
is set, in the case of a normalization filter, weights can be optionally specified by appending :select:minLevel,maxLevel
to the filter specification, which sets positive weights across all levels ranging from minLevel
through maxLevel
, and a zero weight for other levels.
--repeat
/ -r
--verbose
/ -v
--output-folder
/ -o
--merit-digits-displayed