HaltonSequence
[this page | pdf]
Function Description
Returns a Halton quasi-random sequence, where p is
the base prime of the sequence, n is the number of elements in the
sequence returned by the function and istart is the point in the
sequence where we start to return elements.
To calculate the j'th element of a Halton sequence
element (with istart = 0) we write j as a number in base p
(p needs to be prime), e.g. 17 base 3 is 122. We then reverse the
sequence and put a decimal point radix p in front of it, so e.g. the
17'th element of the Halton sequence in base 3 is 0.221 base 3, i.e. (2*9 + 2*3
+ 1*1)/27. Different (positive) values of istart offset the elements of
the sequence that are returned by the function.
Halton sequences are examples of quasi-random, i.e. low
discrepancy, sequences that are more evenly spread out in multiple dimensions
(if, for Halton sequences, the base primes are different for the different
dimensions) than would be the case if elements of the sequences are chosen
‘randomly’, e.g. using merely the MnRand
function (or the equivalent MnProbDistRand with DistributionName
= “uniform” and ParamValues = {0, 1}). Every time the number of digits
in j increases, j’s digits become a factor of p more
finely meshed.
This added smoothness means that in some circumstances they offer
better convergence properties than basic Monte Carlo simulation techniques but at
the cost of it then being less clear what these convergence properties are.
This function does not trap for whether p is prime.
NAVIGATION LINKS
Contents | Prev | Next
Links to:
-
Interactively run function
-
Interactive instructions
-
Example calculation
-
Output type / Parameter details
-
Illustrative spreadsheet
-
Other Statistical functions
-
Computation units used
Note: If you use any Nematrian web service either programmatically or interactively then you will be deemed to have agreed to the Nematrian website License Agreement