This patch release improves reliability by fixing some pesky bugs.
See below for details.
We encourage all Analytica users to install this patch release.
It is available free if you have Analytica 3.1. If you have an older
release, such as 3.0, you may want to consider upgrading now to
3.1. You can upgrade free if you have current maintenance and technical
support (MTS). If your MTS has expired (12 months from purchase)
you can renew it here.
To install
the Patch
If you have Release 3.1, installing 3.1.1 is simple:
Double-click the downloaded
file. It will replace your current 3.1 version, and use your
existing License number, activating the same Edition you already
have.
Defects
fixed in 3.1.1
Distributions
with uncertain parameters
Given uncertain parameter to a normal,
lognormal, or uniform distribution, when using Latin hypercube sampling,
the resulting sample has the correct distribution, but it may have
incorrect correlation with the uncertain parameter(s). For example:
With M := Normal(4, 1) and X := Normal(M, 1), the
distribution of X is correct, but Correlation(X, M)
is wrong. Note that this bug does not affect distributions
with nonprobabilistic parameters or other distribution forms. Nor
does it occur with simple Monte Carlo simulation.
With uncertain parameters to Beta and
Poisson distributions the computation time and memory usage are
large proportional to the square of sample size.
Duplicating
a module
After duplicating
(or copying and pasting) a module, sometimes a variable in the new
module refers to a variable in the original module instead of its
copy in the new module. Sometimes when moving a node into a module,
it creates a "ghost" duplicate node with the same identifier.
Table results
unaligned
A defective
font file on some computers causes column headings in result tables
to be unaligned with their columns.
Miscellaneous
Sometimes Make Importance will include an uncertain variable
more than once in the list of variables.
Analytica can't save model when model title contains a slash
"/"
If you start the same, very large, Analytica model many times
without closing any copies (ignoring Analytica's advice not to
run the same model more than once at a time), it will eventually
crash.
Clicking a Button with a Script referring to "Self"
causes a crash.
Adding a module with the same name as an existing module causes
a crash.
Sometimes when selecting "Copy table" for a very
large table, Analytica crashes.
Using the NLP (NonLinear Programming) from the Optimizer, it
will crash if the decision node has more indexes than expected.