Floating point computation of cycle time for min-max functions

This paper proposes a floating point variant FPCalcCycleTime of the CalcCycleTime algorithm for computing the cycle time of min-max functions. The computation is done by first computing the spectral radius of the function, and then recursively calling itself to compute the cycle time of the reduced function. The original CalcCycleTime algorithm requires all the parameters to be integers, whereas FPCalcCycleTime is able to deal with floating point parameters. This feature makes FPCalcCycleTime well suited for engineering applications.