It is an interesting idea to use the simulated annealing algorithm. It easy to understand, easy to implement, yet the results are not the best (comparing with the other competitors). I wonder how much of this could be improved by tuning the algorithm's properties. This is probably the biggest weakness that the tuning of so many variables will be hard. Since the transformation is fairly simple it could be easily done in plain Java which would likely result in much better execution times.
However, both the report and the solution should really be polished quite a bit. Especially the report.
I would be wondering what is the overhead of using ATL and what is the expressiveness gain of ATL comparing to plain Java.
minor comments:
abstract is missing a reference
"A local optima ..." this sentence does not make sense
"This allows it to ..." missing dot at the end of the sentence
there are number of references that are not connected to the text
it would be nice to provide a way how the solution can be run on multiple models instead of changing the source
It is an interesting idea to use the simulated annealing algorithm. It easy to understand, easy to implement, yet the results are not the best (comparing with the other competitors). I wonder how much of this could be improved by tuning the algorithm's properties. This is probably the biggest weakness that the tuning of so many variables will be hard. Since the transformation is fairly simple it could be easily done in plain Java which would likely result in much better execution times.
However, both the report and the solution should really be polished quite a bit. Especially the report. I would be wondering what is the overhead of using ATL and what is the expressiveness gain of ATL comparing to plain Java.
minor comments:
EvaluationSheet-ATL.xlsx