Reduction benchmarking

This benchmark aims to compare 4 reduction algorithms for nfa.

The compared algorithms are: Hopcroft, Brzozowski, Simulation, Residual. They are implemented in Mata (https://github.com/VeriFIT/mata)

The main metrics are speed, number of states and number of transitions in the resulting automaton

Instalation

Firstly, benchmarks need to be compiled

cd src
make

How to evaluate

I am using pycobench to evaluate the algortihm. To evaluate speed:

./pycobench -c input/bench-reduction.yaml -m [algorithm] -o results/output.out < input/single-automata.input

instead of [algorithm] write the name of the algorithm you want to benchmark

How to parse the output

Pycobench output may not be in optimal format. Parsing the output:

cd results
../pyco_proc --csv output.out > result.csv

Visualising the output

cd visual python3 graph.py [path/to/file] [path/to/file] [states | transitions | time]