parallelize fit_scan_sequence

[dschick]

fit_scan_sequence() can take rather long calculational times which is why, parallelization might be worth thinkt of here, in case the iterations are independent from each other.

[dschick]

it could be possible to use multiprocessing for it, e.g. with a minimal example such as:

with Pool(2*cpu_count()-1) as pool:
    out = pool.starmap(parallel_fitting, (data))
    pool.close()
    pool.join()