closes #75 with PR #100
Add working implementation of parallelized SDP using multiprocessing

small benchmark:

• the computation might be a bit slower on very small instances
• the computational gain in the other cases (discretized state and control spaces size > 100) is between 30 to 50% with 2 processes
• there is a small computational gain even with 1 process for "regular" and large instances

TODO: update doc with frapac help