The existing brute force t-balance test is quite expensive,
especially for large t.
We have implemented a new test, also using brute force but that
performs much better.
However it has a drawback as it might run out of memory space.
For this reason the choice is left to the user as to which
implementation to use for the t-balance test.
The functions where this choice is made possible are
IsDesign, IsSteiner, IsBalanced, Design
as well as the Design<> constructor.
The choice is made using the optional parameter Al
which can be set to the values
- "NoOrbits"
This is the existing implementation of
the brute force test and is the default setting.
- "Orbits"
This is the existing implementation of a test
that uses the orbits of t-sets
under the automorphism group of the incidence structure under
consideration.
This is much faster than "NoOrbits" for some cases,
but slower for others.
- "FastBalanceTest"
This is the new implementation of
the brute force test.
Its usage is
recommended as it significantly outperforms the two previous tests.
These later tests should only be used in the case
where using "FastBalanceTest" would not be possible
due to shortage of memory space.
Typically this would happen for large t and large incidence
structures.
