Applying probability theory for the quality assessment of a wildfire spread prediction framework based on genetic algorithms