sr_errorfuncs.py 文件源码

def __call__(self, individual):
        try:
            # Transform the tree expression in a callable function
            func = toolbox.compile(expr=individual)

            error = numpy.fabs(func(*self.inVarValues) - self.targetVarValues) /  self.targetVarValues
            sumerror = numpy.sum(error)

            return (sumerror,)
        except NameError as ne:
            print ne
            print ne.message
        except Exception as e:
            return (sys.float_info.max,)

#This is a basic error function that finds the maximum of all the relative errors.
#Note that your data set should not contain any 0 values.  That will cause a "divide by zero" error.
#At initialization time, it gets the data to compare against.
#
# The data is a list of lists, the labels give the names of interesting data.
# The config file defines which data lists are of use.
# All data lists are expected to be of equla length.  Repeated values are perfectly OK.
# The config file defines a set of "inVars" which are the input variables.  (e.g. rho, T)
# it also defines a single "targetVar" which is the array of function values.