Fix issue #1

[huyenemma]

The custom transpiler applies some matrix optimization (written in Rewrite_rules.md) from the IBM transpiled circuit:

• Using Pass_manager() with some transformation passes after IBM transpile function.
• Looping running the pass_manager() until there is no improvement.

Already tested the performance and the size of the new transpiler.

TODO: add transformation passes that take benefits from ion-trapped quantum computer (all-to-all connectivity)