postproc: add fine-tuning options for determinization

* spot/twaalgos/postproc.cc, spot/twaalgos/postproc.hh:
Add options det-scc, det-simul, det-stutter.
* bin/spot-x.cc: Document them.
* doc/org/ltl2tgba.org: Illustrate one of them and link
to the spot-x man page.
This commit is contained in:
Alexandre Duret-Lutz 2016-02-10 23:04:43 +01:00
parent 8feab7e2bb
commit b066c6f3f2
4 changed files with 68 additions and 27 deletions

View file

@ -729,7 +729,7 @@ expectations.
deterministic TGBA exists.
#+NAME: ltl2tgba-fga
#+BEGIN_SRC sh :results verbatim :exports none
#+BEGIN_SRC sh :results verbatim :exports code
ltl2tgba "FGa" -D -d.a
#+END_SRC
@ -743,7 +743,7 @@ ltl2tgba "FGa" -D -d.a
But with =--generic=, =ltl2tgba= will output the following Rabin automaton:
#+NAME: ltl2tgba-fga-D
#+BEGIN_SRC sh :results verbatim :exports none
#+BEGIN_SRC sh :results verbatim :exports code
ltl2tgba "FGa" -G -D -d.a
#+END_SRC
@ -758,6 +758,40 @@ Note that determinization algorithm implemented actually outputs
parity acceptance, but =Fin(0)&Inf(1)= can be interpreted either as
=Rabin 1= or =parity min odd 2=.
The [[./man/spot-x.7.html][=spot-x=]](7) man page lists a few =-x= options (=det-scc=,
=det-simul=, =det-stutter=) of the determinization algorithm that are
enable by default, but that you may want to disable for experimental
purpose.
For instance the following deterministic automaton
#+NAME: ltl2tgba-det1
#+BEGIN_SRC sh :results verbatim :exports code
ltl2tgba "F(a W FGb)" -G -D -d.a
#+END_SRC
#+BEGIN_SRC dot :file ltl2tgba-det1.png :cmdline -Tpng :var txt=ltl2tgba-det1 :exports results
$txt
#+END_SRC
#+RESULTS:
[[file:ltl2tgba-det1.png]]
would be larger if SCC-based optimizations were disabled:
#+NAME: ltl2tgba-det2
#+BEGIN_SRC sh :results verbatim :exports code
ltl2tgba "F(a W FGb)" -x '!det-scc' -G -D -d.a
#+END_SRC
#+BEGIN_SRC dot :file ltl2tgba-det2.png :cmdline -Tpng :var txt=ltl2tgba-det2 :exports results
$txt
#+END_SRC
#+RESULTS:
[[file:ltl2tgba-det2.png]]
* Translating multiple formulas for statistics
If multiple formulas are given to =ltl2tgba=, the corresponding